ANTIMICROBIAL STEWARDSHIP IN AUSTRALIAN VETERINARY PRACTICES

by

Laura Yvonne Hardefeldt

ORCID ID: http://orcid.org/0000-0001-5780-7567

This thesis is submitted for the Doctorate of Philosophy in Veterinary Medicine. Faculty of Veterinary and Agricultural Sciences & National Centre for Antimicrobial Stewardship December, 2017

This thesis is being submitted in total fulfilment of the Doctorate of Philosophy in Veterinary Medicine.

i Abstract Antimicrobial use by the veterinary profession has been coming under increasing scrutiny by medical, public health and government officials as the threat of antimicrobial resistance becomes increasingly clear. The World Health Organisation has described antimicrobial resistance as one of the major public health challenges of our time. It is clear that at least some drug-resistant pathogens have evolved under selective pressure from antimicrobial use in agriculture and may be contributing significantly to resistance in clinical setting. Antimicrobial stewardship is the selection of the most appropriate antimicrobial for a given disease in a given animal, with the aim of reducing the risk of adverse effects in that animal, and reducing the likelihood of developing resistance on an individual level, on a farm level and on a national level. Currently none of the core elements of antimicrobial stewardship are widely available for veterinarians in Australia, and there is very sparse data available on which to base an antimicrobial stewardship program. This research project aims to address this paucity of data. A range of research methods were used to assess detailed antimicrobial use by veterinarians in Australia and the enablers and barriers to antimicrobial stewardship. These included quantitative methods such as surveys and analysis of pet insurance data, and qualitative methods such as interviews and focus groups. While antimicrobials with low importance rating were predominately used in all species, under-dosing and inappropriate timing of antimicrobial therapy were common particularly in horses and cattle. Few veterinary practices in Australia had antimicrobial stewardship policies in place, or were using antimicrobial use guidelines. The key barriers to implementing antimicrobial stewardship programs were a lack of antimicrobial stewardship governance structures, client expectations and competition between practices, the cost of microbiological testing, and a lack of access to education, training and antimicrobial stewardship resources. The enablers were concern for the role of veterinary antimicrobial use in development of antimicrobial resistance in humans, a sense of pride in , the firstly, service provided, and preparedness to change prescribing practices. This research culminated in the development secondly, of a proposed antimicrobial stewardship policy thirdly, and procedure documents, to enable veterinarians to institute antimicrobial stewardship programs that suit their individual practice requirements. However, it is likely that governance changes will be necessary to compel veterinary practice owners to implement antimicrobial stewardship on a large scale.

ii Declaration

This thesis comprises only the original work toward the Doctorate of Philosophy in Veterinary Medicine except where indicated in the preface.

Due acknowledgement has been made in the text to all other material used.

The thesis is fewer than the maximum word limit in length, exclusive of tables, maps, bibliographies and appendices as approved by the Research Higher Degrees Committee.

Signed

Laura Hardefeldt

iii Preface All of the work presented henceforth was approved by the University of Melbourne’s Human Research Ethics Committee.

Chapter 1 is currently submitted for publication to Lancet Infectious Diseases: LY Hardefeldt, J Selinger, MA Stevenson, JR Gilkerson, H Crabb, H Billman-Jacobe, K Thursky, KE Bailey, M Awad and GF Browning. Population wide assessment of antimicrobial use in companion animals using a novel data source – a cohort study using pet insurance data. This work was a collaboration with PetSure®. Data compilation was completed by PetSure® data analysts. All major areas of concept formation, data analysis and manuscript composition were completed by L Hardefeldt. Early stages of concept formation and contributions to manuscript edits provided by supervisory authors (GF Browning, MA Stevenson, JR Gilkerson, H Crabb, H Billman Jacobe, K Thursky and KE Bailey) and contributions to manuscript edits provided by PetSure® collaborators prior to submission (J Selinger and M Awad).

Chapter 2 has been published by the Journal of Veterinary Internal Medicine on 17/5/17: LY Hardefeldt, S Holloway, DJ Trott, M Shipstone, VR Barrs, R Malik, M Burrows, S Armstrong, GF Browning and M Stevenson. Antimicrobial Prescribing in Dogs and Cats in Australia: Results of the Australasian Infectious Disease Advisory Panel Survey. J Vet Intern Med 2017; 31(4): 1100-7 This work was a collaboration with the Australasian Infectious Disease Advisory Group (Holloway S, Trott DJ, M. Shipstone, V. R. Barrs, R. Malik, M. Burrows, S. Armstrong) who designed and distributed the questionnaire. All data analysis and the manuscript composition were completed by L Hardefeldt. G. F. Browning and M. Stevenson were the supervisory authors on this project and were involved in project conception and manuscript composition.

Chapter 3 has been published by Veterinary Microbiology on 22/3/17: LY Hardefeldt, GF Browning, K Thursky, JR Gilkerson, H Billman-Jacobe, MA Stevenson and KE Bailey. Antimicrobials used for surgical prophylaxis by companion animal veterinarians in Australia. Vet Microbiol 2017; 203: 301-7. All major areas of concept formation, data collection, data analysis and manuscript composition were completed by L Hardefeldt. Early stages of concept formation and contributions to manuscript edits provided by supervisory authors (Browning GF, Thursky K, JR Gilkerson, H Billman-Jacobe, and MA Stevenson). KE Bailey assisted with data collection, early concept formation and editing of the manuscript.

Chapter 4 has been published online by the Equine Veterinary Journal on 10/7/17: LY Hardefeldt, GF Browning, K Thursky, J. R. Gilkerson, H. Billman-Jacobe, M. A. Stevenson and K. E. Bailey. Antimicrobials used for surgical prophylaxis by equine veterinary practitioners in Australia. Equine Vet J 2017]. All major areas of concept formation, data collection, data analysis and manuscript composition were completed by L Hardefeldt. Early stages of concept formation and contributions to manuscript edits provided by supervisory authors (Browning GF, Thursky K, JR Gilkerson, H Billman-Jacobe, and MA Stevenson). KE Bailey early

iv concept formation, assisted with data collection and contributed to manuscript edits.

Chapter 5 has been published by the Veterinary Record on 19/10/17: LY Hardefeldt, GF Browning, K Thursky, JR Gilkerson, H Billman-Jacobe, MA Stevenson and KE Bailey. A Cross-sectional Study of Antimicrobials used for Surgical Prophylaxis by Bovine Veterinary Practitioners in Australia. Vet Record 2017. All major areas of concept formation, data collection, data analysis and manuscript composition were completed by L Hardefeldt. Early stages of concept formation and contributions to manuscript edits provided by supervisory authors (GF Browning, K Thursky, JR Gilkerson, H Billman-Jacobe, and MA Stevenson). KE Bailey early concept formation, assisted with data collection and contributed to manuscript edits.

Chapter 6 has been accepted for publication by the Australian Veterinary Journal on 22/11/17: LY Hardefeldt, GF Browning, K Thursky, JR Gilkerson, H Billman-Jacobe, MA Stevenson and KE Bailey. Antimicrobial labelling in Australia: a threat to antimicrobial stewardship. AVJ 2017. All major areas of concept formation, data collection, data analysis and manuscript composition were completed by L Hardefeldt. Early stages of concept formation and contributions to manuscript edits provided by supervisory authors (GF Browning, K Thursky, JR Gilkerson, H Billman-Jacobe, and MA Stevenson). KE Bailey early concept formation, assisted with data collection and contributed to manuscript edits.

Chapter 7 is in revision following peer review by the Journal of Veterinary Internal Medicine: LY Hardefeldt, GF Browning, K Thursky, JR Gilkerson, H Billman-Jacobe, MA Stevenson and KE Bailey. The enablers to, and barriers of, implementing antimicrobial stewardship programs in veterinary practices. J Vet Intern Med 2017. All major areas of concept formation, data collection, data analysis and manuscript composition were completed by L Hardefeldt. Early stages of concept formation and contributions to manuscript edits provided by supervisory authors (GF Browning, K Thursky, JR Gilkerson, H Billman-Jacobe, and MA Stevenson) and KE Bailey. GF Browning also assisted with data analysis.

Chapter 8 has been accepted for publication by the Australian Veterinary Journal on 31/10/17: LY Hardefeldt, M Marenda, H Crabb, MA Stevenson, JR Gilkerson, H Billman-Jacobe, Browning GF. Antimicrobial susceptibility testing by the Australian Veterinary Diagnostic Laboratories. AVJ 2017. All major areas of concept formation, data collection, data analysis and manuscript composition were completed by L Hardefeldt. Early stages of concept formation and contributions to manuscript edits provided by supervisory authors (GF Browning, H Crabb, M Marenda, JR Gilkerson, H Billman-Jacobe, MA Stevenson).

v This research was funded by the National Health and Medical Research Council through the Centres of Research Excellence programme, grant no. 1079625. L Hardefeldt was a recipient of an Australian Postgraduate Award scholarship.

vi Acknowledgements

I would like to thank all the veterinarians who participated in this work, and the Australian Veterinary Association for their assistance in recruiting participants through-out. I would also like to acknowledge Agriculture Victoria for their support of the prescribing guidelines and surveys, and for promoting many of the resources.

Thanks are due to my supervisors Professor Glenn Browning, Dr. Helen Billman- Jacobe, Professor Mark Stevenson and Professor Karin Thursky, and the chair of my committee, Professor James Gilkerson. Thanks for giving me such a wonderful opportunity to work with you all, and on this rewarding topic! Also Dr. Kirsten Bailey and Dr. Helen Crabb for all your help.

To my parents, Marie and Kris, for always believing in me and dad for reading through many lists of diseases! I am also grateful to my siblings who have supported me along the way.

Finally, I am profoundly grateful to my wonderful partner, Jeremy, for your unwavering support.

vii Table of Contents

Abstract ...... ii Declaration ...... iii Preface ...... iv Acknowledgements ...... vii List of tables and figures ...... 11 Introduction ...... 14 Chapter 1: ...... 15 Literature review ...... 15 Antimicrobial resistance of animal origin as a public health risk ...... 16 Zoonotic antimicrobial resistant bacteria ...... 16 Antimicrobial resistance in food ...... 17 Antimicrobial resistance surveillance...... 17 Antimicrobial Use in Animals in Australia ...... 19 International Assessment of Veterinary Antimicrobial Use ...... 20 Antimicrobial Stewardship in Veterinary Practices ...... 22 Development of Guidelines for Antimicrobial Use in Veterinary Practice ...... 23 Generic guidelines ...... 23 Disease specific guidelines ...... 23 Successes in Antimicrobial Stewardship in Human Medicine and their Relationship with Veterinary Medicine ...... 23 Education ...... 24 Audit and Feedback ...... 25 Decision support ...... 26 Chapter 2: ...... 38 USING BIG DATA TO INTERROGATE ANTIMICROBIAL USE IN COMPANION ANIMALS IN THE COMMUNITY ...... 38 Population wide assessment of antimicrobial use in companion animals using a novel data source – a cohort study using pet insurance data ...... 39 Abstract ...... 39 Introduction ...... 39 Materials and methods ...... 40 Results ...... 41 Discussion ...... 46 References ...... 48 Chapter 3: ...... 51 ASSESSMENT OF HISTORIC ANTIMICROBIAL PRESCRIBING PATTERNS BY VETERINARIANS IN COMPANION ANIMAL PRACTICE ...... 51 Antimicrobial prescribing in dogs and cats in Australia: results of the Australasian Infectious Disease Advisory Panel survey 2010 ...... 52 Abstract ...... 52 Introduction ...... 52 Methods ...... 53 Results ...... 54 References ...... 58

viii Chapter 4: ...... 60 DETAILED ANTIMICROBIAL USE BY COMPANION ANIMAL VETERINARIANS . 60 Antimicrobials Used for Prophylaxis in Dog and Cat Surgical Scenarios by Veterinarians in Australia ...... 61 Abstract ...... 61 Introduction ...... 61 Methods ...... 62 Results ...... 63 Discussion ...... 66 References ...... 67 Chapter 5: ...... 68 DETAILED ANTIMICROBIAL USE BY EQUINE VETERINARIANS ...... 68 Antimicrobials used for Surgical Prophylaxis by Equine Veterinary Practitioners in Australia ...... 69 Abstract ...... 69 Introduction ...... 69 Materials and methods ...... 69 Results ...... 70 Discussion ...... 72 References ...... 75 Chapter 6: ...... 77 DETAILED ANTIMICROBIAL USE BY BOVINE VETERINARIANS ...... 77 A Cross-Sectional Study of Antimicrobials used for Surgical Prophylaxis by Bovine Veterinary Practitioners in Australia ...... 78 Abstract ...... 78 Introduction ...... 78 Materials and methods ...... 79 Results ...... 79 Discussion ...... 81 References ...... 83 Chapter 7: ...... 84 EFFECT OF HISTORIC LABELLING OF ANTIMICROBIALS ON ANTIMICROBIAL STEWARDSHIP IN VETERINARY PRACTICE...... 84 Antimicrobial labelling in Australia: a threat to antimicrobial stewardship? ...... 85 Abstract ...... 85 Introduction ...... 85 Discussion ...... 88 References ...... 90 Chapter 8: ...... 92 ENABLERS TO, AND BARRIERS OF, ANTIMICROBIAL STEWARDSHIP IN VETERINARY PRACTICE ...... 92 The barriers to, and enablers of, implementing antimicrobial stewardship programs in veterinary practices ...... 93 Abstract ...... 93 Introduction ...... 93 Materials and methods ...... 94 Results ...... 95 Discussion ...... 101

ix References ...... 104 Chapter 9: ...... 107 THE ROLE OF VETERINARY DIAGNOSTIC LABORATORIES IN PROMOTING ANTIMICROBIAL STEWARDSHIP ...... 107 Antimicrobial susceptibility testing methods used by the Australian Veterinary Diagnostic Laboratories ...... 108 Abstract ...... 108 Introduction ...... 108 Methods ...... 109 Results ...... 110 Discussion ...... 114 References ...... 116 Chapter 10: ...... 118 General Discussion ...... 118 Assessing rates of antimicrobial use in veterinary medicine ...... 119 Causes of inappropriate antimicrobial use by veterinarians ...... 120 Enablers and barriers to implementing antimicrobial stewardship ...... 121 Resource development ...... 122 References ...... 123 Summary ...... 125 Appendix 1 ...... 126 Companion animal poster...... 126 Large animal flip book ...... 127 Appendix 2 ...... 135 Antimicrobial stewardship policy ...... 135 Antimicrobial stewardship procedure template ...... 136

x BACK TO TABLE OF List of tables and figures CONTENTS

Chapter 1: Table 1. A logistic regression model of risk factors for total antimicrobial usage.

Table 2. A logistic regression model of risk factors for critically important antimicrobial usage.

Figure 1. Number of antimicrobial prescriptions for the most common conditions affecting dogs between 2013 and 2017.

Figure 2. Number of antimicrobial prescriptions for the most common conditions affecting cats between 2013 and 2017.

Figure 3. Proportion of use of critically important antimicrobials over time for dogs and cats.

Figure 4. Monthly incidence rate of antimicrobial prescribing in dogs between 2013 and 2017.

Figure 5. Monthly incidence rate of antimicrobial prescribing in cats between 2013 and 2017.

Chapter 2: Table 1. Overall frequency of antibiotic use across medical, surgical and dermatological scenarios posed in this survey.

Figure 1. Agreement with AIDAP guidelines for choice of empirical or antimicrobial therapy guided by C & S, choice of drug and duration of therapy, and overall agreement with the guidelines for treatment of (a) gingivitis, (b) pyothorax, (c) acute cystitis and (d) peritonitis. White columns indicate treatment choices for cats and black columns indicate the treatment choices for dogs.

Chapter 3: Table 1. Guidelines for antimicrobial prophylaxis in specific surgical scenarios.

Table 2. Estimated regression coefficients and their standard errors from a logistic regression model of risk factors for antimicrobial usage compliance.

Figure 1. Frequency of reported antimicrobial usage for seven clinical scenarios.

Figure 2. Proportions of antimicrobial usage, by class, for surgical scenarios. * HIRA: High importance rating antimicrobial. ** LIRA: Low importance rating antimicrobial

Figure 3. Proportions of veterinarians reporting optimal or suboptimal compliance with AIDAP and BSAVA guidelines for prophylactic therapy for different surgical scenarios.

Figure 4. Proportions of veterinarians reporting compliance with guidelines for choice of antimicrobial drug, the timing and route of administration or the duration of therapy for different surgical scenarios.

10 Chapter 4: Table 1. Estimated regression coefficients and their standard errors from a logistic regression model of risk factors for compliance with guidelines for prophylactic antimicrobial usage in surgery.

Figure 1. Frequency of antimicrobial usage for surgical prophylaxis in seven scenarios.

Figure 2. Overall proportions of antimicrobials reported as being used in surgical prophylaxis. *TMS: Trimethoprim sulphonamide

Figure 3. Antimicrobials used for prophylaxis in each of the surgical scenarios. *LIRA: Low importance rating antimicrobial

Figure 4. Proportions of veterinarians reporting optimal and suboptimal compliance with BEVA guidelines for prophylactic antimicrobial use in different surgical scenarios. Suboptimal compliance reflects appropriate drug choice but inappropriate doses or timing of antimicrobial administration to allow for adequate serum antimicrobial concentrations at the time of surgery, or a duration of therapy that was not compliant with guidelines.

Figure 5. Proportions of veterinarians reporting sub-optimal compliance with antimicrobial prophylaxis guidelines evaluated by factor.

Chapter 5: Table 1. Survey respondent demographics

Figure 1. Frequency of prophylactic antimicrobial usage in cattle for different surgical scenarios.

Figure 2. Overall proportion of antimicrobials used for surgical prophylaxis across all scenarios. *TMS: Trimethoprim sulphonamide

Figure 3. Proportions of different classes of antimicrobials for surgical prophylaxis in specific scenarios. *LIRA: Other low importance rating antimicrobials

Figure 4. Proportions of respondents indicating differing durations of antimicrobial prophylactic therapy for specific surgical scenarios.

Chapter 6: Figure 1. Doses of antimicrobials reported by equine practitioners for (a) procaine penicillin (1000IU/kg) intramuscularly, (b) trimethoprim/sulphonamide orally and (c) gentamicin (mg/kg) intravenously. Arrows indicate recommended dose.

Figure 2. Doses of antimicrobials reported by companion animal practitioners for (a) amoxycillin (mg/kg) orally and (b) amoxycillin/clavulanate (mg/kg) subcutaneously. Arrows indicate recommended dose.

Table 1. Recommended doses and inter-dosing intervals for antimicrobials in horses.

Table 2. Recommended doses and inter-dosing intervals for antimicrobials in cattle.

11 Chapter 7: Figure 1. Qualitative study logistics.

Figure 2. Proportion of survey respondents indicating how much antimicrobial use by individuals, and by the profession, contributes to the overall burden of AMR.

Figure 3. Frequency of use of antimicrobials with high-importance rating. HIRA; high-importance rating antimicrobials.

Table 1. Demographics of survey respondents and interview participants compared to national veterinary workforce.

Table 2. Summary of major barriers and enablers and implementing AMS programs in veterinary practices.

Table 3. Summary of recommendations to facilitate the establishment of AMS programs in veterinary practices.

Chapter 8: Table 1. Most commonly isolated pathogens for each animal species included in the survey.

Table 2. Antimicrobials commonly included in susceptibility testing for different pathogens.

12 BACK TO TABLE OF Introduction CONTENTS

Antimicrobial stewardship (AMS) is the movement to improve antimicrobial prescribing and reduce the pressure on the development of antimicrobial resistance (AMR). Antimicrobial stewardship programs for veterinarians have had minimal development globally in comparison to the medical profession. In Australia there have been few studies into antimicrobial use by veterinarians, and no assessment of appropriateness of use. By assessing appropriate antimicrobial use we can identify areas of prescribing, or sectors of the profession, that are underperforming and target these in AMS programs.

The aims of this thesis were to: 1. Investigate the rate of antimicrobial exposure in a cohort of companion animals. 2. Report on historic antimicrobial use in companion animals and investigate reasons for inappropriate antimicrobial use. 3. Investigate detailed antimicrobial usage, compliance with guidelines (where possible), and need for additional guidelines and antimicrobial stewardship in companion animal, bovine and equine practice. 4. Identify the enablers to, and barriers of, antimicrobial stewardship in veterinary practices. 5. To develop and economically viable, effective and adaptable antimicrobial stewardship package for veterinary practices.

The rate of antimicrobial prescribing in the companion animal community is discussed in chapter 1 and this puts into context the findings in the following 2 chapters that investigates detailed antimicrobial use, and reasons for inappropriate antimicrobial use, in companion animals veterinary practice (chapter 2 and 3). Although rates of antimicrobial use in the large animal population still represent a gap in the literature, this thesis investigated detailed antimicrobial use, and reasons for inappropriate antimicrobial use in equine veterinary practice (chapter 4) and in bovine veterinary practice (chapter 5). Specifically the role of antimicrobial drug labelling is addressed in chapter 6.

From the results in the first 6 chapters, this thesis argues that the there is a need for antimicrobial stewardship, and given this finding, went on to investigate the enablers to, and barriers of, antimicrobial stewardship in this population (chapter 7 and 8). The culmination of this research is a proposal of an antimicrobial stewardship policy and a procedure document that can be adapted by veterinarians to suit their individual practice needs (appendix 2).

13 BACK TO TABLE OF CONTENTS

Chapter 1:

Literature review

15 Antimicrobial resistance of animal origin as a public health risk

In 1917 scientists at the Rockefeller Institute added Sulfanilamide to quinine derivatives in an effort to improve bacterial killing, and the first clinical use of Sulphonamides is reported in 19331. In 1928 Alexander Fleming discovered penicillin and since these times there have been dramatic improvements in health care as a result of the use of antimicrobials. However, bacteria have a natural capacity to resist the actions of many antimicrobial agents2. With this added evolutionary pressure to adapt to this new environment, bacteria have developed drug resistance, and increasingly multi-drug resistance (MDR). These MDR pathogens have been selected for in veterinary and medical practice since the introduction of antimicrobials almost 90 years ago, but they have assumed growing global significance in recent years as the rate of development of novel antimicrobials slows. Although a national surveillance system for AMR has recently been developed in Australia3, data on morbidity and mortality associated with AMR is still lacking. However in the United States of America, the Centres for Disease Control and Prevention estimates that more than two million people are affected by AMR pathogens each year, with at least 23,000 deaths resulting annually4. A review in the United Kingdom (UK) in 2014 estimated the global economic cost of MDR infections and found that, if current trends continue, by 2050 around 10 million people may die each year as a result of AMR. In addition, gross domestic product would decrease by 2-3.5%, costing the world economy approximately US$140 trillion5. Social and health costs were not considered and could significantly increase this figure.

Zoonotic antimicrobial resistant bacteria The transfer of multidrug resistant pathogens between food animals and humans, and between companion animals and humans, is contested. Multi-drug resistance pathogens, such as methicillin resistant Staphylococcus aureus (MRSA), have been identified in food animals6-9 and companion animals10-18, as well as their owners18- 20, farm workers9, 20, 21, veterinarians17, 18, 20, 22-24, and others involved in animal industries15, 20, 22, 25. The MRSA strains that are carried in companion animals are generally similar to human strains8, 26, 27, but it is unclear whether humans or animals were the initial source of these strains. Many authors claim that companion animals are a reservoir of resistance, but these claims are unreferenced or not supported by the evidence from their research10, 14, 28, and MRSA are generally not considered to be a commensal organism of dogs and cats. In livestock, however, the strains of MRSA that have been detected are not those typically associated with humans and these have therefore been designated livestock-associated MRSA (LA-MRSA)8. These strains have been isolated at higher rates in those people with high levels of contact with livestock8, leading to the conclusion that animals are a source of LA-MRSA for those working in livestock- associated professions25. In addition, these strains have been shown to spread between household members, indicating that person-to-person spread of LA- MRSA can occur29. Similarly, Staphylococcus pseudintermedius is not usually associated with the respiratory microbiome of humans. Methicillin-resistant S. pseudintermedius (MRSP) has been isolated from owners of colonised animals, indicating that zoonotic transmission of MDR pathogens does occur30, 31. It is

16 therefore clear that direct, or indirect, contact with animals can result in acquisition of MDR pathogens.

Antimicrobial resistance in food However, from a public health standpoint, the more important route of potential transmission of AMR is through food. MDR bacteria can be selected in animals in response to antimicrobial use7, 16, 32-39, as they are in humans, and food may be contaminated with AMR bacteria, or with antimicrobial resistance genes. In the United States of America (USA) extended-spectrum cephalosporin resistant Salmonella enterica serovar Heidelberg have been linked to chicken meat40. However, multidrug resistant isolates have been found more commonly in human isolates than in food isolates6, 41, suggesting that food is not the most likely source of AMR in this pathogen. Retail chicken meat in the USA has higher levels of ceftriaxone-resistant Salmonella Typhimurium than cultures taken from chickens at slaughter, suggesting confounding factors may be contributing to the presence of AMR pathogens in meat42. Globally, the degree to which MDR pathogens in animals contribute to the reservoir of resistance, and can then be transmitted to people through food, or direct or indirect contact, has yet to be elucidated. Many studies find that the rate of MDR pathogens on farms and in meat is very low6, 43-46, except on pig21, 44 and chicken47, 48 farms, where MDR bacteria appear to have a higher prevalence. However, there is a widely circulated belief by some influential individuals that the volume of antimicrobial use in food animals is correlated with the amount of AMR seen in human populations and poses a public health risk49-54. This view has been strengthened by such phenomena as the rapid increase in ciprofloxacin-resistant campylobacteriosis following introduction of fluoroquinolones into broiler chickens in the USA55 and the decline in extended spectrum cephalosporin resistance in Salmonella isolates following the voluntary withdrawal of third generation cephalosporins from the poultry market in Canada56. Consistent with this are the low levels of fluoroquinolone resistance seen in Salmonella isolates in Australia3, where fluoroquinolones have never been registered for use in food producing animals. While such conclusions are tempting to draw, this seems likely to be an overly simplistic view of a situation into which many interventions are being introduced. No multivariate analyses to investigate such theories can be found in the current literature.

Antimicrobial resistance surveillance In Australia, comprehensive national surveillance has not yet been conducted, so the extent of AMR in the community is unknown. Information is restricted to that published in the literature and to reports from the Australian Commission on Safety and Quality in Health Care, which released the Antimicrobial Use and Resistance in Australia (AURA)3 report in 2016 and the Critical Antimicrobial Resistance (CARAlert)57 report in 2017. CARAlert uses the antimicrobial importance rating system defined by the Australian Strategic Technical Advisory Group on AMR58. This system has been used to define critically important (also called high-importance rating) throughout this thesis. These reports bring together many Australian surveillance programs for human disease and describe trends in AMR in important human pathogens, but AMR in animal pathogens is not included these reports and the bacteria included are restricted to clinical isolates obtained from diagnostic laboratories. Carbapenemase-producing Enterobacteriaceae (CPE)

17 and azithromycin non-susceptible Neisseria gonorrhoeae account for most of the critical antimicrobial resistances in 2016 and 2017 in Australia57. N. gonorrhoeae is not a pathogen of animals and does not have zooanthroponotic potential. While companion animals may develop infections59 with, and may serve as a reservoir for, CPE60, currently surveillance of healthy companion animals is not performed in Australia, so the risk of zoonotic transfer of these pathogens to animal owners cannot be estimated.

In Australia, there have been limited surveillance data for AMR in bacteria of animal origin even though this is recognised as important in the National Implementation Plan for tackling AMR61. The Department of Agriculture and Water Resources, from November 2003 to June 2004, undertook a pilot surveillance program. Samples from 204 cattle, 200 pigs and 303 chickens were collected from 31 abattoirs. From these, 645 Escherichia coli, 547 presumed Enterococcus species, and 133 Campylobacter species were isolated. There were no MDR E. coli isolated from cattle, but 63.2% of isolates collected from pigs were resistant to 2 or more antimicrobials and 26.4% were resistant to 4 or more antimicrobials. Of the chicken isolates, 34.6% were resistant to 2 or more antimicrobials and 2.6% resistant to 4 or more antimicrobials. This trend was also seen with Enterococcus species, with cattle having a lower prevalence of resistant isolates than chickens, and pigs having the highest prevalence (9.5%, 45.9% and 93.3% of cattle, chicken and pig isolates resistant to erythromycin, and 9.5%, 28.7% and 43.3% of cattle, chicken and pig isolates resistant to virginiamycin, respectively) and resistance to 2 antimicrobials being absent in isolates from cattle, but present in 11.5% of chicken isolates and 46.7% of pig isolates. Campylobacter species were only isolated from chickens, with 19.7% resistant to tetracyclines, 9.8% resistant to erythromycin and 1.5% resistant to both antimicrobials. With the exception of Enterococcus faecium, in which virginiamycin resistance was common, there was no, or very low levels of, resistance to antimicrobials of importance in human medicine44. A similar study was completed by the CSIRO and the NSW Department of Primary Industries in 2013, but only samples were only obtained from cattle and Salmonella species and E. coli isolated62. Salmonella from cattle had low rates of resistance, with the prevalence not exceeding 3.7% for any one antimicrobial and 91.5% of isolates from beef cattle, and all veal and dairy cattle isolates, sensitive to all the antimicrobials tested except florfenicol. AMR was low in E. coli isolates, with resistance only to florfenicol and tetracyclines seen in beef and dairy cattle isolates. In veal cattle isolates, resistance was detected for amoxicillin- clavulanic acid, gentamicin, kanamycin and ceftiofur, although at very low levels (1.1%, 0.6%, 1.1% and 0.6% respectively)62. A second AMR surveillance pilot study, conducted in 2013, collected sensitivity data on close to 2600 E. coli and coagulase positive Staphylococcus species (predominately S. aureus and S. pseudintermedius) from all 22 veterinary diagnostic laboratories in Australia. A very low incidence of extended-spectrum cephalosporin resistance and fluoroquinolone resistance was found in livestock63. Coagulase positive Staphylococcus species from companion animals (dogs, cats and horses) had frequencies of resistance of 11.8% for S. pseudintermedius and 12.8% for S. aureus64. There was no carbapenem resistance in any E. coli isolates. There was also a low prevalence of oxacillin and cefoxitin resistance in S. aureus isolates, with this occurring seen only in isolates from dogs and horses65, 66.

18 The NSW Department of Primary Industries funded a surveillance program on Salmonella enterica isolates from cattle from 2007 to 201167. This study also found a low prevalence of resistance, with 66.1% of isolates remaining susceptible to all antimicrobials tested and no resistance to fluoroquinolones or third-generation cephalosporins detected in the isolates67. Other smaller studies investigating AMR in Australia have had similar results, with very low resistance in Histophilus somni isolated from cattle68 and Salmonella isolates from layer chickens and eggs69.

Comprehensive AMR prevalence surveys have not been conducted on companion animals, however two studies have investigated the similarities between human and companion animal MDR E. coli, with both revealing considerable similarities19, 70. An analysis of Australian MRSA isolates from animals and veterinarians suggests both zoonotic and zooanthroponotic transfer71. MRSP has been isolated and characterised from canine pyoderma cases64, 72 and the prevalence of carriage of resistance has been estimated in Victoria (0.4% of isolates showed methicillin resistance)73, but prevalence in the wider Australian canine population has not been established.

In summary, levels of AMR in bacteria isolated from cattle and poultry have been, and continue to be, very low. However, the emergence of MRSP suggests that antimicrobial use in dogs is contributing to the emergence of antimicrobial resistance. Isolation of other resistant pathogens is complicated by possible zooanthroponotic transfer from animal owners or nosocomial infection from veterinary hospitals or veterinarians. Similar confounding factors affect the interpretation of reports of antimicrobial resistance in equines. Resistance levels in pigs appear higher, and may suggest antimicrobial use in this species is contributing to the emergence of AMR, however further evaluation is required to elucidate the true association.

Antimicrobial Use in Animals in Australia

There is considerable evidence that antimicrobial use drives AMR. Therefore, the amount of antimicrobial use should be optimised for animal health outcomes to minimise the risk of AMR development. Australia has very conservative legislation restricting antimicrobial use in veterinary practice, particularly in food producing species. Fluoroquinolones are banned for use in food-producing animals and 4th generation cephalosporins have not been registered for use in animals in Australia. The volume of antimicrobials sold for agricultural and veterinary use is monitored by Australian Pesticides and Veterinary Medicines Authority (APVMA). Their report from 2014, for antimicrobial sales from 2005-2010, gives the total tonnes of active constituent of antimicrobials sold in 2009-2010 at 661.2 tonnes74. Approximately 98% of all antimicrobials sold for animal use were sold for food animals and about 43% of these were sold for therapeutic or prophylactic use, with the remaining 56% being used predominately as coccidiostats, with only 4- 7% for use as growth promotion. Of the antimicrobials sold for food animals, 49% were for poultry (predominately coccidiostats), 36% for pigs and 15% for cattle and sheep74. However, outside of the use of antimicrobials as coccidiostats, there is no information on what antimicrobials are used, and for what indication in each species.

19 There have been two Australian industry surveys investigating antimicrobial use in specific animal industries. A report by AgVet Projects, with data supplied by Dairy Australia, describes antimicrobial use sold by dairy cattle veterinary practices in the 2012-13 financial year75. Penicillins and sulphonamides predominated (45% and 40% of all sales, respectively), with tetracyclines (7%), aminoglycosides (2%), cephalosporins (3%) and macrolides (3%) making up the remainder75. The pig industry was surveyed in 2009, with self-reported antimicrobial use data reported. Surveys were completed by 51% of the pig producers in Australia, with producers reporting the most commonly used antimicrobials to be penicillins, tetracyclines, sulphonamides and macrolides. Ceftiofur usage was reported by 24.8%, and aminoglycoside use by 51.7%, of herds76.

Three surveys have been performed to investigate antimicrobial use in small animal practice in Australia. In 1997 a survey was performed by the University of Sydney to investigate systemic antibacterial drug use in dogs77. Survey respondents were asked about the patterns of use of various systemic antibacterial drugs and their approach to 9 specific medical scenarios. Penicillins and cephalosporins were most commonly used, with amoxicillin-clavulanate the most frequently reported antimicrobial agent. Empirical antimicrobial therapy was selected in the vast majority of acute medical conditions (76-94% of cases), and was frequently used in chronic conditions (15-50% of cases)77. This survey was repeated in 2017 with Victorian veterinarians only78. Empirical antimicrobial therapy remained commonly selected for acute medical conditions (42-92% of cases), whereas culture and susceptibility testing appeared to have increased in chronic conditions (78-98% of chronic cases having samples submitted for culture and susceptibility testing). While penicillins and cephalosporins remained the most commonly prescribed antimicrobials, there were changes in the prescribing practices for a few discrete scenarios78. A third, similar study was performed by the Australian Infectious Diseases Advisory Panel (AIDAP) in 2010, but was not published and the results of this survey now form a part of this thesis. Independent investigation of antimicrobial use for food animals and the broader companion animal population is lacking and this represents a gap in the literature for Australian data.

International Assessment of Veterinary Antimicrobial Use

Many methods have been used internationally in attempts to assess veterinary antimicrobial use. As with Australia, many countries have data limited to the gross tonnage of antimicrobials sold to the agricultural sector. Europe monitors veterinary antimicrobial sales through the European Medicines Agency, which is tasked with evaluating and supervising the medicines used for human and veterinary patients79. The agency launched the European Surveillance of Veterinary Antimicrobial Consumption (ESVAC) in 2009. ESVAC has produced 5 reports since its formation, with the most recent report presenting trends in antimicrobial usage by class, amongst other things, across 26 countries79. The data collected cover approximately 95% of the food-producing animal populations in the European Union/European Economic Area. However, these data only extend to food-producing animals. Any drugs sold in tablet form are excluded from the

20 report as these are presumed to be for use in small animal patients. Notably, a legal basis exists in 24 of the 26 countries contributing to the data to allow ESVAC to request data on sales or prescriptions from importers, wholesalers and end-users, although end-user data is available in all countries. Data are corrected to a population unit to allow comparisons between countries. Where data have been collected for 4 or more consecutive years, a valid baseline is regarded to have been established and trends can be investigated79. This system is the most complete assessment of veterinary antimicrobial usage at the time of writing.

In South Africa, a similar, but not continuous, process was undertaken from 2002 to 2004 to provide a point prevalence estimate of antimicrobial usage across the veterinary industry by obtaining sales data from 8 of the 25 veterinary pharmaceutical companies80. This allowed in-feed, water, and parenteral antimicrobial usage to be estimated, but the indication for use was not able to be deduced, so appropriateness cannot be evaluated. Small individual studies have investigated antimicrobial use in small animal patients in the United Kingdom81, 82 and use of critically important antimicrobials in horses in the United Kingdom83 and the United States84. Many of these studies used data mining from electronic medical records to quantify the total number of prescriptions and administrations of antimicrobials at a practice level. As all consultations were examined, a denominator can be deduced and the prevalence of antimicrobial use reported. It was not possible to differentiate between new and repeat prescriptions and administrations, so incidence could not be reported. In addition, the indication for antimicrobial use was not able to be obtained using this approach, as there was a lack of use of standardised nomenclature to record clinical diagnoses81. This is likely to be a consistent limitation of research that employs data mining, but useful information about gross antimicrobial usage and antimicrobial usage by class can be obtained on a practice-by-practice basis, which can contribute to the overall assessment of veterinary antimicrobial usage.

Other individual studies have also been conducted to investigate antimicrobial use in specific areas of veterinary medicine. These have mostly been in university practices, so the representativeness of these data across the entire veterinary profession is limited85-87. In Ontario, Canada, a survey was undertaken to investigate oral and parenteral antimicrobial use in dogs and cats in general practice88. Veterinarians kept a log-book one day per month over a 12-month period, recording antimicrobial use for up to a maximum of 5 eligible patients that were presented on each of their journal days. This prospective research approach allowed elimination of recall bias that may occur with research that relies on a veterinarian’s memory or speculation about hypothetical situations, methods that have been used in several studies in developed nations77, 82, 89-91. A similar study was also conducted to investigate antimicrobial use in cattle in Finland, but veterinarians only collected data for one week92. While these surveys provide insights into antimicrobial use at a specific time-point, they have limited usefulness in antimicrobial use surveillance.

21 Antimicrobial Stewardship in Veterinary Practices

Antimicrobial stewardship is the use of the correct antimicrobial, at the correct dose, route and duration, for the correct bacterial infection at the correct time93. In addition, antimicrobial stewardship is the restraint from using antimicrobials when they are not indicated. The primary goal of antimicrobial stewardship is to optimise clinical outcomes, while minimising the unintended consequence of antimicrobial use, such as toxicity, the selection of pathogenic organisms and the emergence of antimicrobial resistance94. Antimicrobial stewardship programs (ASPs) for veterinarians have had minimal development globally in comparison to the medical profession, where uptake in hospitals in the developed world is common. In the medical profession, ASPs are generally a set of interventions at a hospital level aimed at improving compliance with guidelines, reducing antimicrobial use, reducing resistance and improving clinical outcomes at a local level95. In contrast, antimicrobial stewardship in veterinary medicine has generally been associated with regulation aimed at restricting antimicrobial use in the veterinary profession, and limiting residues in food produced from animals. There are several opinion-style articles in the literature93, 96, 97, but there are no reports of implementation of a hospital-style ASP in private veterinary practices.

The only component of an ASP in animal health that can be assessed objectively is drug restriction, as has occurred in European countries since the 1970s, when there was a general ban on the use of tetracyclines, penicillins and streptomycin for growth promotion98. Since that time, several Scandinavian countries have taken a more conservative approach and lobbied the European Union (EU) to ban use of all antimicrobials for growth promotion, a restriction that came into effect in the EU in 200698. Some countries have also implemented further restrictions, such as limiting veterinary profits from antimicrobial sales (Denmark, Finland) and ceasing the use of all general prophylactic antimicrobials for animals (Sweden, Denmark). Reports in the literature assessing the impact of these measures on the overall burden of AMR and AMR infections in humans are absent. There are, however, reports assessing the change in usage and the effects on productivity in selected countries. Antimicrobial consumption by Danish pig farms from 1992 to 2008 was evaluated alongside a sample of productivity data from 10% of pig farmers. Use of antimicrobials had reduced by more than 50% over the study period, with an improvement in productivity suggesting that the ban on antimicrobials for growth promotion was not adversely affecting the long-term productivity of the pig industry in Denmark99. A greater target for reduction has been set in the Netherlands, where the goal is to reduce antimicrobial use in farm animals by 70% between 2007 and 201554. A report in 2012, indicated that use had been reduced by 56%54. The effect of this reduction on AMR had not been assessed in these early stages. Reducing antimicrobial use did coincide with a reduction in the burden of AMR in pigs100, 101 and poultry100 over the long-term, but not over the short term102, in European countries. This is important for the veterinary and agricultural sectors, but the importance for the medical sector has yet to be elucidated.

22 Development of Guidelines for Antimicrobial Use in Veterinary Practice

Generic guidelines Generic antimicrobial guidelines are commonly adopted by veterinary association to promote judicious antimicrobial use by members. The Australian Veterinary Association has a code of practice for prescription and use of products which contain antimicrobial agents103 and a fact-sheet on prescribing veterinary antibiotics104. The effectiveness of generic guidelines, that promote the principles of responsible prescribing practices, in improving antimicrobial stewardship has not been investigated.

Disease specific guidelines The British Small Animal Veterinary Association and the British Equine Veterinary Association have released antimicrobial policy poster templates for veterinary practices to guide first line antimicrobial selection, and alternatives to these first line drugs, for a range of common indications105, 106. The equine document also includes a guide to drug doses and dosing frequencies106. Unfortunately, there has been no assessment of the effectiveness of these guidelines in improving antimicrobial prescribing in practices and significant differences exist in drug availability and indications between Britain and Australia.

The Danish Small Animal Veterinary Association has produced comprehensive guidelines for antimicrobial use in a range of medical conditions107 and the Canadian Veterinary Medical Association has produced similar guidelines for the prudent use of antimicrobials in beef cattle, dairy cattle and pigs108. Sweden has comprehensive antimicrobial guidelines109, and the International Society for Companion Animal Infectious Diseases has released comprehensive guidelines for individual syndromes (urinary tract110, superficial bacterial folliculitis111 and respiratory tract disease112), however these tend to be focused on companion animals in the United States of America, with uncertain relevance for the Australian companion animal population. The Australasian Infectious Diseases Advisory Panel (AIDAP) have produced limited guidelines for small animal practice in Australia113, but these guidelines are sponsored by a pharmaceutical company and the impact of this on veterinarians’ attitudes to use of this document is unknown. It does appear that promotion of the one of the products of this pharmaceutical company does occur frequently in the document, potentially affecting perceptions of the validity of the guidelines. In all cases, no evaluation of the uptake, or compliance, with these guidelines could be found in the literature. Evidence-based and independent guidelines are needed for veterinarians in Australia to guide antimicrobial therapy for a range of species.

Successes in Antimicrobial Stewardship in Human Medicine and their Relationship with Veterinary Medicine

The drivers of ASPs have typically been the need to prevent future outbreaks of nosocomial MDR infections in hospitals, with secondary goal of reducing hospital costs without adversely affecting quality of care94. Increasingly, particularly in Australia, ASPs are mandated by regulatory bodies that fund human hospitals114.

23 While regulation does assist in ensuring widespread implementation of ASPs, the consistent financial benefits from effectively established ASPs have meant that many of these programs can be self-supporting in both large115-121 and small hospitals122, 123. There is strong evidence in the literature that planned interventions can change prescribing practices and can control infectious disease outcomes116, 124-126.

Components of an antimicrobial stewardship program in medical practice can be restrictive or persuasive. Restrictive interventions reduce the freedom of prescribers to select some antimicrobials. In medical practice, restriction is implemented at a local level (within hospitals) and is therefore different to the restrictive interventions that have occurred in Europe where restriction comes in the form of legislation. Persuasive interventions are aimed at behavioural change and are focused on addressing predisposing factors, through measures such as such as practitioner education and practice guidelines, reinforcing factors, through measures such as providing professional or clinical champions, and audit and feedback, or enabling, factors through measures such as patient education and decision support127. Many medical ASP programs have elements of both restriction and persuasion, and in a Cochrane systematic review neither was found to be more successful than the other over the long term95. However, while implementing local restriction of selected antimicrobials, with review and advice by specialist infectious disease physicians, may be feasible in many human hospitals115, 121, 128, 129, it is very difficult to implement such measures in veterinary practices, not least because the specialty of infectious diseases does not exist in veterinary medicine. In addition, many veterinary practitioners are in solo practice; for example approximately 45% of veterinary practitioners work alone130, so in their setting there is no local expert from whom to gain permission. Even if a body existed from whom a veterinary practitioner could seek permission to use a restricted drug, the geographical isolation of many veterinarians and the need for round the clock access to support would make any such program difficult to implement. Persuasive interventions, therefore, are likely to form the basis of veterinary ASPs. In the medical setting, persuasive interventions have been intensively studied in a range of settings. For this review, only those settings with the most relevance to veterinary practice in Australia will be examined, with brief mention of the systematic reviews and meta-analyses that have been performed across sectors. Those settings that are considered are general medical practice, hospitals in low- to-middle income countries and other resource poor settings. Although some studies attempt to investigate a single intervention, multiple interventions are more commonly investigated so these are also discussed below.

Education

Education is included as a core element of many antimicrobial stewardship programs94, 131, 132 and has been identified by a citizen jury in Australia as an area for policy intervention133. A review by Soumerai et. al.134, in 1989, examined primary care educational interventions aimed at improving prescribing. Printed educational materials, educational outreach, group education and feedback to prescribers were all found to have convincing evidence for effectiveness. However, in a systematic review, printed educational interventions were found to have

24 minimal benefit on both professional practice and healthcare outcomes135. However, in this meta-analysis, the interventions were used alone and compared to no intervention, and the effectiveness of printed educational materials as part of a multifaceted approach was not investigated. Academic detailing is another form of education that has been widely investigated in primary care practices. Academic detailing is non-commercial face-to-face educational outreach that is typically provided by trained health care professionals. This type of education has been shown to reduce antimicrobial prescriptions136-138 and reduce inappropriate antimicrobial prescriptions136 in some studies, but not in others139-142. Standardised educational seminars have also been investigated and found to reduce antimicrobial prescriptions in the short143 and long143, 144 term. Training in communication has been shown to reduce prescriptions of antimicrobials in primary care settings145, 146, but training to improve probabilistic disease judgments had no impact on antimicrobial prescribing, despite improved estimation of disease probability147. These types of face-to-face educational interventions have been shown to reduce antimicrobial prescribing in conjunction with other interventions148 and to improve the effectiveness of other interventions, such as audit and feedback, in low resource settings149, 150, but not in others151. In addition, Finkelstein et al, in a 2008 study of primary care practices in Massachusetts, found that the benefits of a multifaceted educational package, including prescribing feedback, were sectoral, with the more robust impact among Medicaid-insured children and for broad-spectrum agents152. This highlights the challenging nature of this area, where the socioeconomic status of the patient or client, along with a multitude of other factors, are likely to affect decision making by doctors, and the type, delivery and content of the educational intervention affect the efficacy of such programs. In addition, a single intervention is less likely to produce significant results, with more complex interventions or an entire stewardship package found to be more successful, albeit in the hospital, not primary care, sector. The veterinary sector will be vastly different again from any of these settings, and the gap in published literature on veterinary medicine about this subject will only be filled once ASPs are trialled and evaluated in veterinary practice.

Audit and Feedback

Audit and feedback can be prospective or retrospective. Prospective audit with intervention and feedback typically occurs in hospitals, where an infectious disease clinician or infectious disease clinical pharmacist interacts with the prescriber and offers feedback on appropriate antimicrobial use153. This form of audit and feedback has been associated with appropriate antimicrobial selection and with lower rates of resistance153, but the relevance to veterinary practice is limited for the reasons stated above. Retrospective audit, on the other hand, may be possible in veterinary practice. This type of audit has been used and evaluated in low resource and primary medical practice in Australia154 and overseas141, 155, 156. Again, however, the results have been mixed, with some studies finding a beneficial effect154-156 and others failing to do so141. This diversity itself has also been investigated with attention focused on the theoretical and conceptual bases underlying audit and feedback. According to Control theory, behaviour change is most likely if feedback is accompanied by comparison with a behavioural target

25 and action plans157. Colquhoun et al. found that the explicit use of theory in audit and feedback studies was rare and not consistent, possibly explaining the variability158. The use of theory can help to understand both the mechanisms and design of interventions when behaviour change is the ultimate goal159, 160. Any studies performed in veterinary practice should be theoretically sound in order to advance our understanding of this educational method. Retrospective audit may be possible in veterinary practice with tools such as VetCompass161, which uses data mining of electronic medical records and practice management software, and will be capable of detecting differences in gross antimicrobial use and changes in antimicrobial use between practices and interventions. Other methods that have been used in medical practice include repeated surveys of treatment preferences for different syndromes and diseases, with feedback on appropriateness154. Making practitioners accountable for their prescribing decisions was identified by a citizen jury in Australia as being an area where policy intervention may reduce antimicrobial use133. However, while this type of audit is also possible in veterinary medicine, it is unlikely to be sustainable, as developing, performing and analysing this type of survey is expensive and time-consuming, and practitioners are likely to develop survey fatigue.

Decision support

Decision support was introduced into medical practice to combat the limited influence that passive dissemination of guidelines and consensus-derived recommendations were having on behavioural change162, 163. Electronic decision support is used in medical practice to combine guideline recommendations with individual patient data and population statistics, such as local antibiograms, to provide relevant, objective, accurate and up-to-date recommendations for patient therapy164. The recommendation can be empirical or follow provision of results from microbiological testing. A systematic review has investigated the success of electronic decision support systems and overall 54% had a positive impact, with the remainder having no impact. All systems were effective in the management of acute disease, but only 38% improved management of chronic conditions165, once again showing that antimicrobial stewardship is much more complex in primary care than in a hospital setting. In addition to this challenge, for these types of antimicrobial stewardship initiatives to be developed for veterinary practice, the foundational tools would first need to be established, namely guidelines and regional antibiograms. While regional antibiogram publication in veterinary medicine is limited by the willingness of private laboratories to share data, practices may be willing to share resistance patterns in an anonymous manner if a user-friendly system was established. Decision support is typically an add-on to electronic medical record systems in the medical profession166. While these systems are becoming more frequently used in small animal practice, their use in large animal practice is still intermittent. Without widespread use of these systems, decision support software cannot be developed for veterinary medicine.

Consistent with the discussion above, a systematic review of antimicrobial stewardship in outpatient settings in 2015 found medium strength evidence that stewardship programs that incorporate communication skills training and laboratory testing reduced antimicrobial use, but only low-strength evidence that

26 other stewardship interventions were associated with improved prescribing167. How stewardship interventions will perform in veterinary practice is unknown, but the lessons from medical practice, particularly those from primary care and low resource settings, should be heeded in the design of any stewardship package for veterinary practice.

1. Shambaugh GE. History of Sulfonamides. Archives of Otolaryngology - Head and Neck Surgery 1966;83:1-2. 2. D'Costa VM, King CE, Kalan L et al. Antibiotic resistance is ancient. Nature 2011;477:457-461. 3. Australian Commission on Safety and Quality in Health Care. AURA 2016: The first report on antimicrobial use and resistance in human health. ACSQHC, Sydney, 2016. 4. CDC. Antibiotic resistance threats in the United States, 2013. Centres for Disease Control and Prevention, www.cdc.gov, 2013. 5. The Review on Antimicrobial Resistance. Antimicrobial Resistance: Tackling a crisis for the health and wealth of nations. In: O'Neill J, editor, www.amr- review.org, 2014. 6. Melo DB, Menezes AP, Reis JN, Guimaraes AG. Antimicrobial resistance and genetic diversity of Escherichia coli isolated from humans and foods. Braz J Microbiol 2015;46:1165-1170. 7. Jiang X, Yang H, Dettman B, Doyle MP. Analysis of fecal microbial flora for antibiotic resistance in ceftiofur-treated calves. Foodborne Pathog Dis 2006;3:355-365. 8. Pantosti A. Methicillin-resistant Staphylococcus aureus associated with animals and its relevance to human health. Front Microbiol 2012;3:127. 9. Oppliger A, Moreillon P, Charriere N et al. Antimicrobial resistance of Staphylococcus aureus strains acquired by pig farmers from pigs. Appl Environ Microbiol 2012;78:8010-8014. 10. Davis JA, Jackson CR, Fedorka-Cray PJ et al. Carriage of methicillin-resistant Staphylococci by healthy companion animals in the US. Lett Appl Microbiol 2014;59:1-8. 11. Wan MT, Fu SY, Lo YP et al. Heterogeneity and phylogenetic relationships of community-associated methicillin-sensitive/resistant Staphylococcus aureus isolates in healthy dogs, cats and their owners. J Appl Microbiol 2012;112:205-213. 12. Lilenbaum W, Nunes EL, Azeredo MA. Prevalence and antimicrobial susceptibility of Staphylococci isolated from the skin surface of clinically normal cats. Lett Appl Microbiol 1998;27:224-228. 13. Loeffler A, Pfeiffer DU, Lindsay JA, Soares Magalhaes RJ, Lloyd DH. Prevalence of and risk factors for MRSA carriage in companion animals: a survey of dogs, cats and horses. Epidemiology and infection 2011;139:1019- 1028. 14. Boost MV, O'Donoghue MM, James A. Prevalence of Staphylococcus aureus carriage among dogs and their owners. Epidemiology and infection 2008;136:953-964.

27 15. Grinberg A, Kingsbury DD, Gibson IR et al. Clinically overt infections with methicillin-resistant Staphylococcus aureus in animals in New Zealand: a pilot study. N Z Vet J 2008;56:237-242. 16. Rentala M, Lahti E, Kuhalampi J et al. Antimicrobial resistance in Staphlococcus spp., Escherichia coli and Enterococcus spp. in dogs given antibiotics for chronic dermatological disorders compared with non-treated control dogs. Acta Vet Scand 2004;45:37-45. 17. Schwaber MJ, Navon-Venezia S, Masarwa S et al. Clonal transmission of a rare methicillin-resistant Staphylococcus aureus genotype between horses and staff at a veterinary teaching hospital. Vet Microbiol 2013;162:907-911. 18. Weese JS, Dick H, Willey BM et al. Suspected transmission of methicillin- resistant Staphylococcus aureus between domestic pets and humans in veterinary clinics and in the household. Vet Microbiol 2006;115:148-155. 19. Platell JL, Cobbold RN, Johnson JR et al. Commonality among fluoroquinolone-resistant sequence type ST131 extraintestinal Escherichia coli isolates from humans and companion animals in Australia. Antimicrob Agents Chemother 2011;55:3782-3787. 20. Liu W, Liu Z, Yao Z et al. The prevalence and influencing factors of methicillin-resistant Staphylococcus aureus carriage in people in contact with livestock: a systematic review. Am J Infect Control 2015;43:469-475. 21. Dohmen W, Bonten MJ, Bos ME et al. Carriage of extended-spectrum beta- lactamases in pig farmers is associated with occurrence in pigs. Clin Microbiol Infect 2015;21:917-923. 22. Ishihara K, Saito M, Shimokubo N et al. Methicillin-resistant Staphylococcus aureus carriage among veterinary staff and dogs in private veterinary clinics in Hokkaido, Japan. Microbiol Immunol 2014;58:149-154. 23. Paterson GK, Harrison EM, Craven EF et al. Incidence and characterisation of methicillin-resistant Staphylococcus aureus (MRSA) from nasal colonisation in participants attending a cattle veterinary conference in the UK. PLoS One 2013;8:e68463. 24. Paul NC, Moodley A, Ghibaudo G, Guardabassi L. Carriage of methicillin- resistant Staphylococcus pseudintermedius in small animal veterinarians: indirect evidence of zoonotic transmission. Zoonoses Public Health 2011;58:533-539. 25. Boost M, Ho J, Guardabassi L, O'Donoghue M. Colonization of butchers with livestock-associated methicillin-resistant Staphylococcus aureus. Zoonoses Public Health 2013;60:572-576. 26. Vincze S, Stamm I, Kopp PA et al. Alarming proportions of methicillin- resistant Staphylococcus aureus (MRSA) in wound samples from companion animals, Germany 2010-2012. PLoS One 2014;9:e85656. 27. Ishihara K, Saito M, Shimokubo N et al. Epidemiological analysis of methicillin-resistant Staphylococcus aureus carriage among veterinary staff of companion animals in Japan. J Vet Med Sci 2014;76:1627-1629. 28. Couto N, Monchique C, Belas A et al. Trends and molecular mechanisms of antimicrobial resistance in clinical Staphylococci isolated from companion animals over a 16 year period. J Antimicrob Chemother 2016;71:1479-1487. 29. Bosch T, Verkade E, van Luit M et al. Transmission and persistence of livestock-associated methicillin-resistant Staphylococcus aureus among

28 veterinarians and their household members. Appl Environ Microbiol 2015;81:124-129. 30. Ishihara K, Shimokubo N, Sakagami A et al. Occurrence and molecular characteristics of methicillin-resistant Staphylococcus aureus and methicillin-resistant Staphylococcus pseudintermedius in an academic veterinary hospital. Appl Environ Microbiol 2010;76:5165-5174. 31. Walther B, Hermes J, Cuny C et al. Sharing more than friendship--nasal colonization with coagulase-positive Staphylococci (CPS) and co-habitation aspects of dogs and their owners. PLoS One 2012;7:e35197. 32. Lowrance TC, Loneragan GH, Kunze DJ et al. Changes in antimicrobial susceptibility in a population of Escherichia coli isolated from feedlot cattle administered ceftiofur crystalline-free acid. Am J Vet Res 2007;68:501-507. 33. Schmidt JW, Griffin D, Kuehn LA, Brichta-Harhay DM. Influence of therapeutic ceftiofur treatments of feedlot cattle on fecal and hide prevalences of commensal Escherichia coli resistant to expanded-spectrum cephalosporins, and molecular characterization of resistant isolates. Appl Environ Microbiol 2013;79:2273-2283. 34. Alali WQ, Scott HM, Norby B, Gebreyes W, Loneragan GH. Quantification of the bla(CMY-2) in feces from beef feedlot cattle administered three different doses of ceftiofur in a longitudinal controlled field trial. Foodborne Pathog Dis 2009;6:917-924. 35. Leite-Martins LR, Mahu MI, Costa AL et al. Prevalence of antimicrobial resistance in enteric Escherichia coli from domestic pets and assessment of associated risk markers using a generalized linear mixed model. Prev Vet Med 2014;117:28-39. 36. Kanwar N, Scott HM, Norby B et al. Effects of ceftiofur and chlortetracycline treatment strategies on antimicrobial susceptibility and on tet(A), tet(B), and bla CMY-2 resistance genes among E. coli isolated from the feces of feedlot cattle. PLoS One 2013;8:e80575. 37. Ahmed MO, Clegg PD, Williams NJ, Baptiste KE, Bennett M. Antimicrobial resistance in equine faecal Escherichia coli isolates from North West England. Ann Clin Microbiol Antimicrob 2010;9:12. 38. Berge AC, Epperson WB, Pritchard RH. Assessing the effect of a single dose florfenicol treatment in feedlot cattle on the antimicrobial resistance patterns in faecal Escherichia coli. Vet Res 2005;36:723-734. 39. Chambers L, Yang Y, Littier H et al. Metagenomic analysis of antibiotic resistance genes in dairy cow feces following therapeutic administration of third generation cephalosporin. PLoS One 2015;10:e0133764. 40. Folster JP, Pecic G, Singh A et al. Characterization of extended-spectrum cephalosporin-resistant Salmonella enterica serovar Heidelberg isolated from food animals, retail meat, and humans in the United States 2009. Foodborne Pathog Dis 2012;9:638-645. 41. Thorsteinsdottir TR, Haraldsson G, Fridriksdottir V, Kristinsson KG, Gunnarsson E. Prevalence and genetic relatedness ofantimicrobial-resistant Escherichia coli isolated from animals, foods and humans in Iceland. Zoonoses and Public Health 2010;57:189-196. 42. Iwamoto M, Reynolds J, Karp BE et al. Ceftriaxone-Resistant Nontyphoidal Salmonella from Humans, Retail Meats, and Food Animals in the United States, 1996-2013. Foodborne Pathog Dis 2017;14:74-83.

29 43. Carson CA, Reid-Smith R, Irwin RJ, Martin WS, McEwen SA. Antimicrobial resistance in generic fecal Escherichia coli from 29 beef farms in Ontario. Can J Vet Res 2008;72:119-128. 44. Department of Agriculture Fisheries and Forestry. Pilot surveillance program for antimicrobial resistance in bacteria of animal origin. http://www.agriculture.gov.au, 2007. 45. Alexander TW, Inglis GD, Yanke LJ et al. Farm-to-fork characterization of Escherichia coli associated with feedlot cattle with a known history of antimicrobial use. Int J Food Microbiol 2010;137:40-48. 46. Schoster A, Staempfli HR, Arroyo LG et al. Longitudinal study of Clostridium difficile and antimicrobial susceptibility of Escherichia coli in healthy horses in a community setting. Vet Microbiol 2012;159:364-370. 47. Sheikh AA, Checkley S, Avery B et al. Antimicrobial resistance and resistance genes in Escherichia coli isolated from retail meat purchased in Alberta, Canada. Foodborne Pathog Dis 2012;9:625-631. 48. Agerso Y, Jensen JD, Hasman H, Pedersen K. Spread of extended spectrum cephalosporinase-producing Escherichia coli clones and plasmids from parent animals to broilers and to broiler meat in a production without use of cephalosporins. Foodborne Pathog Dis 2014;11:740-746. 49. Economou V, Gousia P. Agriculture and food animals as a source of antimicrobial-resistant bacteria. Infect Drug Resist 2015;8:49-61. 50. Verraes C, Van Boxstael S, Van Meervenne E et al. Antimicrobial resistance in the food chain: a review. Int J Environ Res Public Health 2013;10:2643- 2669. 51. Chantziaras I, Boyen F, Callens B, Dewulf J. Correlation between veterinary antimicrobial use and antimicrobial resistance in food-producing animals: a report on seven countries. J Antimicrob Chemother 2014;69:827-834. 52. Collignon P, Aarestrup FM. Extended-spectrum B-lactamases, food and cephalosporin use in food animals. Clin Infect Dis 2007;44:1391-1392. 53. Vieira AR, Collignon P, Aarestrup FM et al. Association between antimicrobial resistance in Escherichia coli isolates from food animals and blood stream isolates from humans in Europe: an ecological study. Foodborne Pathog Dis 2011;8:1295-1301. 54. Speksnijder DC, Mevius DJ, Bruschke CJ, Wagenaar JA. Reduction of veterinary antimicrobial use in the Netherlands. The Dutch success model. Zoonoses Public Health 2015;62 Suppl 1:79-87. 55. Gupta A, Nelson JM, Barrett TJ et al. Antimicrobial resistance among Campylobacter strains, United States, 1997-2001. Emerging Infectious Diseases 2004;10:1102-1109. 56. Dutil L, Irwin R, Finley R et al. Ceftiofur resistance in Salmonella enterica serovar Heidelberg from chicken meat and humans, Canada. Emerg Infect Dis 2010;16:48-54. 57. Australian Commission on Safety and Quality in Health Care. CARAlert first annual report March 2016-March 2017. ACSQHC, Sydney, 2017. 58. Australian Strategic and Technical Advisory Group on Antimicrobial Resistance. Importance rating and summary of antibacterials used in human health in Australia. 2015. Retrieved 5/2/18.

30 59. Gibson JS, Morton JM, Cobbold RN et al. Multidrug-resistant E. coli and Enterobacter extraintestinal infection in 37 dogs. J Vet Intern Med 2008;22:844-850. 60. Shaheen BW, Nayak R, Boothe DM. Emergence of a New Delhi metallo-beta- lactamase (NDM-1)-encoding gene in clinical Escherichia coli isolates recovered from companion animals in the United States. Antimicrob Agents Chemother 2013;57:2902-2903. 61. Commonwealth of Australia. National antimicrobial resistance strategy 2015-2019. http://www.health.gov.au/internet/main/publishing.nsf/Content/1803C4 33C71415CACA257C8400121B1F/$File/amr-strategy-2015-2019.pdf, 2016. 62. Barlow RS, McMillan KE, Duffy LL et al. Prevalence and antimicrobial resistance of Salmonella sp. and Escherichia coli from Australian cattle populations at slaughter. Journal of Food Protection 2015;78:912-920. 63. Abraham S, Jordan D, Wong HS et al. First detection of extended-spectrum cephalosporin- and fluoroquinolone-resistant Escherichia coli in Australian food-producing animals. J Glob Antimicrob Resist 2015;3:273-277. 64. Saputra S, Jordan D, Worthing KA et al. Antimicrobial resistance in coagulase-positive Staphylococci isolated from companion animals in Australia: a one year study. PLoS One 2017;12:e0176379. 65. Shaban RZ, Simon GI, Trott DJ, Turnidge J, Jordan D. Surveillance and reporting of antimicrobial resistance and antibiotic usage in animals and agriculture in Australia. Department of Agriculture, Griffith University, University of Adelaide, www.agriculture.gov.au/publications, 2014. 66. Abraham S, Trott DJ. First national survey of antimicrobial resistance in Escherichia coli and coagulase positive Staphylococcus spp. isolated from clinical infections in animals. AVA conference, Newcastle, 2014. 67. Abraham S, Groves MD, Trott DJ et al. Salmonella enterica isolated from infections in Australian livestock remain susceptible to critical antimicrobials. Int J Antimicrob Agents 2014;43:126-130. 68. Goldspink LK, Mollinger JL, Barnes TS et al. Antimicrobial susceptibility of Histophilus somni isolated from clinically affected cattle in Australia. Vet J 2015;203:239-243. 69. Pande VV, Gole VC, McWhorter AR, Abraham S, Chousalkar KK. Antimicrobial resistance of non-typhoidal Salmonella isolates from egg layer flocks and egg shells. Int J Food Microbiol 2015;203:23-26. 70. Sidjabat HE, Townsend KM, Hanson ND et al. Identification of bla(CMY-7) and associated plasmid-mediated resistance genes in multidrug-resistant Escherichia coli isolated from dogs at a veterinary teaching hospital in Australia. J Antimicrob Chemother 2006;57:840-848. 71. Worthing KA, Abraham S, Pang S et al. Molecular Characterization of Methicillin-Resistant Staphylococcus aureus Isolated from Australian Animals and Veterinarians. Microb Drug Resist 2017;Epub. 72. Siak M, Burrows AK, Coombs GW et al. Characterization of meticillin- resistant and meticillin-susceptible isolates of Staphylococcus pseudintermedius from cases of canine pyoderma in Australia. J Med Microbiol 2014;63:1228-1233.

31 73. Bean DC, Wigmore SM. Carriage rate and antibiotic susceptibility of coagulase-positive Staphylococci isolated from healthy dogs in Victoria, Australia. Aust Vet J 2016;94:456-460. 74. Australian Pesticides and Veterinary Medicines Authority. Quantity of antimicrobial products sold for veterinary use in Australia: July 2005 to June 2010. www.apvma.gov.au, 2014. 75. Greenall R. Antimicrobial use in the Australian dairy industry. AgVet Projects, 2014. 76. Jordan D, Chin JJ, Fahy VA et al. Antimicrobial use in the Australian pig industry: results of a national survey. Aust Vet J 2009;87:222-229. 77. Watson ADJ, Maddison JE. Systemic antibacterial drug use in dogs in Australia. Australian Veterinary Journal 2001;79:740-746. 78. Barzelai ID, Whittem T. Survey of systemic antimicrobial prescribing for dogs by Victorian veterinarians. Aust Vet J 2017;95:375-385. 79. European Medicines Agency. Sales of veterinary antimicrobial agents in 26 EU/EEA countries in 2013: Fifth ESVAC report. www.ema.europa.eu, 2015. 80. Eagar H, Swan G, van Vuuren M. A survey of antimicrobial usage in animals in South Africa with specific reference to food animals. J S Afr Vet Assoc 2012;83:16. 81. Mateus A, Brodbelt DC, Barber N, Stark KD. Antimicrobial usage in dogs and cats in first opinion veterinary practices in the UK. J Small Anim Pract 2011;52:515-521. 82. Mateus AL, Brodbelt DC, Barber N, Stark KD. Qualitative study of factors associated with antimicrobial usage in seven small animal veterinary practices in the UK. Prev Vet Med 2014;117:68-78. 83. Hammersley E, Marshall JF, Parkin TDH, Duz M, Welsh CE. Retrospective study of the use of reserved antimicrobials in first-opinion UK equine veterinary practice. Equine Veterinary Journal, Birmingham, UK, 2016. 84. Welsh CE, Parkin TDH, Duz M, Marshall JF. Use of restricted antimicrobials in US first-opinion equine medicine. Equine Veterinary Journal, Birmingham, UK, 2016. 85. Weese JS. Investigation of antimicrobial use and the impact of antimicrobial use guidelines in a small animal veterinary teaching hospital: 1995-2004. J Am Vet Med A 2006;228:553-558. 86. Rantala M, Holso K, Lillas A, Huovinen P, Kaartinen L. Survey of condition- based prescribing of antimicrobial drugs for dogs at a veterinary teaching hospital. Veterinary Record 2004;155:259-+. 87. Holso K, Rantala M, Lillas A et al. Prescribing antimicrobial agents for dogs and cats via university pharmacies in Finland--patterns and quality of information. Acta Vet Scand 2005;46:87-93. 88. Murphy CP, Reid-Smith RJ, Boerlin P et al. Out-patient antimicrobial drug use in dogs and cats for new disease events from community companion animal practices in Ontario. Canadian Veterinary Journal-Revue Veterinaire Canadienne 2012;53:291-298. 89. Pleydell EJ, Souphavanh K, Hill KE, French NP, Prattley DJ. Descriptive epidemiological study of the use of antimicrobial drugs by companion animal veterinarians in New Zealand. N Z Vet J 2012;60:115-122.

32 Knights CB, Mateus A, Baines SJ. Current British veterinary attitudes to the use of perioperative antimicrobials in small animal surgery. Vet Rec 90. 2012;170:646. De Briyne N, Atkinson J, Pokludova L, Borriello SP. Antibiotics used most commonly to treat animals in Europe. Vet Rec 2014;175:325. 91. Thomson K, Rantala M, Hautala M, Pyorala S, Kaartinen L. Cross-sectional prospective survey to study indication-based usage of antimicrobials in 92. animals: results of use in cattle. BMC Vet Res 2008;4:15. Guardabassi L, Prescott JF. Antimicrobial stewardship in small animal veterinary practice: from theory to practice. Vet Clin North Am Small Anim 93. Pract 2015;45:361-376, vii. Dellit TH, Owens RC, McGowan JE, Jr. et al. Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America guidelines 94. for developing an institutional program to enhance antimicrobial stewardship. Clin Infect Dis 2007;44:159-177. Davey P, Brown E, Charani E et al. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev 95. 2013:CD003543. Turnidge J. Antimicrobial stewardship: what is it, and how does it work? Animal Production Science 2015. 96. Pakyz AL, McDanel J. A US view on antimicrobial stewardship in animal agriculture. Current Treatment Options in Infectious Diseases 97. 2016;8:115-123. Cogliani C, Goosens H, Greko C. Restricting antimicrobial use in food animals: lessons from Europe. Microbe 2011;6:274-279. 98. Aarestrup FM, Jensen VF, Emborg HD, Jacobsen E, Wegener HC. Changes in the use of antimicrobials and the effects on productivity of swine farms in 99. Denmark. American Journal of Veterinary Research 2010;71:726-733. Aarestrup FM, Seyfarth AM, Emborg HD et al. Effect of abolishment of the use of antimicrobial agents for growth promotion on occurrence of 100. antimicrobial resistance in fecal Enterococci from food animals in Denmark. Antimicrob Agents Chemother 2001;45:2054-2059. Agerso Y, Aarestrup FM. Voluntary ban on cephalosporin use in Danish pig production has effectively reduced extended-spectrum cephalosporinase- 101. producing Escherichia coli in slaughter pigs. J Antimicrob Chemother 2013;68:569-572. Pakpour S, Jabaji S, Chenier MR. Frequency of antibiotic resistance in a swine facility 2.5 years after a ban on antibiotics. Microb Ecol 102. 2012;63:41-50. Australian Veterinary Association. Appendix 15 - Australian Veterinary Association (AVA) code of practice for prescription and use of products 103. which contain antimicrobial agents. 2008. Retrieved 26/2/18. Australian veterinary Association. Prescribing veterinary antibiotics. 2013. Retrieved 26/2/18. 104. British Small Animal Veterinary Association. PROTECT. https://www.bsava.com/Resources/Veterinary-resources/PROTECT. 2016. 105. Retrieved 16/2/18. British Equine Veterinary Association. PROTECT ME. 2016. Retrieved. Spohr A, Schjoth B, Wiinberg B et al. Antibiotic use guidelines for 106. companion animal practice. Danish Small Animal Veterinary Association, 107. 33 www.fecava.org/sites/default/files/DSAVA_AntibioticGuidelines%20v1- 1_3(1).pdf, 2009. 108. Blackwell T, McClure JT, Tremblay R et al. Canadian Veterinary Medical Assocation Antimicrobial prudent use guidelines 2008. 2008. 109. Swedish Veterinary Association. Guidelines for the clinical use of antibiotics in the treatment of dogs and cats. 2009. Retrieved 26/2/18. 110. Weese JS, Blondeau JM, Boothe D et al. Antimicrobial use guidelines for treatment of urinary tract disease in dogs and cats: antimicrobial guidelines working group of the international society for companion animal infectious diseases. Vet Med Int 2011;2011:263768. 111. Hillier A, Lloyd DH, Weese JS et al. Guidelines for the diagnosis and antimicrobial therapy of canine superficial bacterial folliculitis (Antimicrobial Guidelines Working Group of the International Society for Companion Animal Infectious Diseases). Vet Dermatol 2014;25:163-175, e142-163. 112. Lappin MR, Blondeau J, Boothe D et al. Antimicrobial use Guidelines for Treatment of Respiratory Tract Disease in Dogs and Cats: Antimicrobial Guidelines Working Group of the International Society for Companion Animal Infectious Diseases. J Vet Intern Med 2017;31:279-294. 113. Holloway S, Trott DJ, Shipstone M et al. Antibiotic prescribing; detailed guidelines. Australasian Infectious Diseases Advisory Panel, http://www.ava.com.au/sites/default/files/AVA_website/pdfs/AIDAP%20 prescribing%20guidelines.pdf, 2013. 114. Australian Commission on Safety and Quality in Health Care. National safety and quality health service standards. ACSQHC, www.safetyandquality.gov.au/wp- content/uploads/2011/09/NSQHSStandards-Sept-2012.pdf., 2011. 115. Schentag JJ, Ballow CH, Fritz AL et al. Changes in antimicrobial agent usage resulting from interactions among clinical pharmacy, the infectious disease division, and the microbiology laboratory. Diagn Microbiol Infect Dis 1993;16:255-264. 116. Carling P, Fung T, Killion A, Terrin N, Barza M. Favorable impact of a multidisciplinary antibiotic management program conducted during 7 years. Infect Control Hosp Epidemiol 2003;24:699-706. 117. Ansari F, Gray K, Nathwani D et al. Outcomes of an intervention to improve hospital antibiotic prescribing: interrupted time series with segmented regression analysis. J Antimicrob Chemother 2003;52:842-848. 118. Lutters M, Harbarth S, Janssens JP et al. Effect of a comprehensive, multidisciplinary, educational program on the use of antibiotics in a geriatric university hospital. J Am Geriatr Soc 2004;52:112-116. 119. Scheckler WE, Bennett JV. Antibiotic usage in seven community hospitals. JAMA 1970;213:264-267. 120. Standiford HC, Chan S, Tripoli M, Weekes E, Forrest GN. Antimicrobial stewardship at a large tertiary care academic medical center: cost analysis before, during, and after a 7-year program. Infect Control Hosp Epidemiol 2012;33:338-345. 121. White AC, Jr., Atmar RL, Wilson J et al. Effects of requiring prior authorization for selected antimicrobials: expenditures, susceptibilities, and clinical outcomes. Clin Infect Dis 1997;25:230-239.

34 122. LaRocco A, Jr. Concurrent antibiotic review programs--a role for infectious diseases specialists at small community hospitals. Clin Infect Dis 2003;37:742-743. 123. Ruttimann S, Keck B, Hartmeier C, Maetzel A, Bucher HC. Long-term antibiotic cost savings from a comprehensive intervention program in a medical department of a university-affiliated teaching hospital. Clin Infect Dis 2004;38:348-356. 124. de Man P, Verhoeven BA, Verbrugh HA, Vos MC, van den Anker JN. An antibiotic policy to prevent emergence of resistant bacilli. Lancet 2000;355:973-978. 125. Singh N, Rogers P, Atwood CW, Wagener MM, Yu VL. Short-course empiric antibiotic therapy for patients with pulmonary infiltrates in the intensive care unit: a proposed solution for indiscriminate antibiotic prescription. Am J Respir Crit Care Med 2000;162:505-511. 126. Bradley SJ, Wilson AL, Allen MC et al. The control of hyperendemic glycopeptide-resistant Enterococcus spp. on a haematology unit by changing antibiotic usage. J Antimicrob Chemother 1999;43:261-266. 127. Green LW, Kreuter MW. Health program planning: an educational and ecological approach. 4th edn. McGraw-Hill, New York, NY, 2005. 128. Medina Presentado JC, Paciel Lopez D, Berro Castiglioni M, Gerez J. Ceftriaxone and ciprofloxacin restriction in an intensive care unit: less incidenceof Acinetobacter spp. and improved susceptibility of Pseudomonas aeruginosa. Rev Panam Salud Publica 2011;30:603-609. 129. Lewis GJ, Fang X, Gooch M, Cook PP. Decreased resistance of Pseudomonas aeruginosa with restriction of ciprofloxacin in a large teaching hospital's intensive care and intermediate care units. Infect Control Hosp Epidemiol 2012;33:368-373. 130. Australian Bureau of Statistics. 8564.0 - Veterinary services, Australia, 1999-2000. www.abs.gov.au, 2001. 131. Public Health England. Start Smart - Then focus: Antimicrobial stewardship toolkit for English hospitals. www.gov.uk/phe, 2015. 132. CDC. Core elements of hospital antibiotic stewardship programs. US Department of Health and Human Services, www.cdc.gov/getsmart/healthcare/implementation/core-elements.html, 2014. 133. Degeling C, Johnson J, Iredell J et al. Assessing the public acceptability of proposed policy interventions to reduce the misuse of antibiotics in Australia: A report on two community juries. Health Expect 2017. 134. Soumerai SB, McLaughlin TJ, Avorn J. Improving drug prescribing in primary care: a critical analysis of the experimental literature. Milbank Q 1989;67:268-317. 135. Giguere A, Legare F, Grimshaw J et al. Printed educational materials: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev 2012;10:CD004398. 136. Seager JM, Howell-Jones RS, Dunstan FD et al. A randomised controlled trial of clinical outreach education to rationalise antibiotic prescribing for acute dental pain in the primary care setting. Br Dent J 2006;201:217-222; discussion 216.

35 137. Gjelstad S, Hoye S, Straand J et al. Improving antibiotic prescribing in acute respiratory tract infections: cluster randomised trial from Norwegian general practice (prescription peer academic detailing (Rx-PAD) study). BMJ 2013;347:f4403. 138. Avorn J, Soumerai SB. Improving drug-therapy decisions through education outreach: a randomised controlled trial of academically based "detailing". NEJM 1983;308:1457-1463. 139. de Burgh S, Mant A, Mattick RP et al. A controlled trial of educational visiting to improve benzodiazepine prescribing in general practice. Aust J Public Health 1995;19:142-148. 140. Enriquez-Puga A, Baker R, Paul S, Villoro-Valdes R. Effect of educational outreach on general practice prescribing of antibiotics and antidepressants: a two-year randomised controlled trial. Scand J Prim Health Care 2009;27:195-201. 141. Mainous AG, 3rd, Hueston WJ, Love MM, Evans ME, Finger R. An evaluation of statewide strategies to reduce antibiotic overuse. Fam Med 2000;32:22- 29. 142. Doyne EO, Alfaro MP, Siegel RM et al. A randomized controlled trial to change antibiotic prescribing patterns in a community. Arch Pediatr Adolesc Med 2004;158:577-583. 143. Le Corvoisier P, Renard V, Roudot-Thoraval F et al. Long-term effects of an educational seminar on antibiotic prescribing by GPs: a randomised controlled trial. Br J Gen Pract 2013;63:e455-464. 144. Welschen I, Kuyvenhoven MM, Hoes AW, Verheij TJ. Effectiveness of a multiple intervention to reduce antibiotic prescribing for respiratory tract symptoms in primary care: randomised controlled trial. BMJ 2004;329:431. 145. Cals JW, Butler CC, Hopstaken RM, Hood K, Dinant GJ. Effect of point of care testing for C reactive protein and training in communication skills on antibiotic use in lower respiratory tract infections: cluster randomised trial. BMJ 2009;338:b1374. 146. Altiner A, Brockmann S, Sielk M et al. Reducing antibiotic prescriptions for acute cough by motivating GPs to change their attitudes to communication and empowering patients: a cluster-randomized intervention study. J Antimicrob Chemother 2007;60:638-644. 147. Poses RM, Cebul RD, Wigton RS. You can lead a horse to water-improving physicians’ knowledge of probabilities may not affect their decisions. Med Decis Making 1995;15:65-75. 148. Finkelstein JA, Davis RL, Dowell SF et al. Reducing antibiotic use in children: a randomized trial in 12 practices. Pediatrics 2001;108:U113-U119. 149. Awad AI, Eltayeb IB, Baraka OZ. Changing antibiotics prescribing practices in health centers of Khartoum State, Sudan. Eur J Clin Pharmacol 2006;62:135-142. 150. Gerber JS, Prasad PA, Fiks AG et al. Effect of an outpatient antimicrobial stewardship intervention on broad-spectrum antibiotic prescribing by primary care pediatricians: a randomized trial. JAMA 2013;309:2345-2352. 151. Naughton C, Feely J, Bennett K. A RCT evaluating the effectiveness and cost- effectiveness of academic detailing versus postal prescribing feedback in changing GP antibiotic prescribing. J Eval Clin Pract 2009;15:807-812.

36 152. Finkelstein JA, Huang SS, Kleinman K et al. Impact of a 16-community trial to promote judicious antibiotic use in Massachusetts. Pediatrics 2008;121:e15-23. 153. DiazGranados CA. Prospective audit for antimicrobial stewardship in intensive care: impact on resistance and clinical outcomes. Am J Infect Control 2012;40:526-529. 154. Zwar N, Wolk J, Gordon J, Sanson-Fisher R, Kehoe L. Influencing antibiotic prescribing in general practice: a trial of prescriber feedback and management guidelines. Fam Pract 1999;16:495-500. 155. Monette J, Miller MA, Monette M et al. Effect of an educational intervention on optimizing antibiotic prescribing in long-term care facilities. J Am Geriatr Soc 2007;55:1231-1235. 156. Ayieko P, Ntoburi S, Wagai J et al. A multifaceted intervention to implement guidelines and improve admission paediatric care in Kenyan district hospitals: a cluster randomised trial. PLoS Med 2011;8:e1001018. 157. Carver CS, Scheier MF. Control theory: a useful conceptual framework for personality-social, clinical, and health psychology. Psychol Bull 1982;92:111-135. 158. Colquhoun HL, Brehaut JC, Sales A et al. A systematic review of the use of theory in randomized controlled trials of audit and feedback. Implementation Science 2013;8:66-74. 159. Gardner B, Whittington C, McAteer J, Eccles MP, Michie S. Using theory to synthesise evidence from behaviour change interventions: the example of audit and feedback. Soc Sci Med 2010;70:1618-1625. 160. Michie S, Fixsen D, Grimshaw JM, Eccles MP. Specifying and reporting complex behaviour change interventions: the need for a scientific method. Implement Sci 2009;4:40. 161. Kearsley-Fleet L, O'Neill DG, Volk HA, Church DB, Brodbelt DC. Prevalence and risk factors for canine epilepsy of unknown origin in the UK. Vet Rec 2013;172:338. 162. Benson T. Why general practitioners use computers and hospital doctors do not - Part 2: scalability. British Medical Journal 2002;325:1090-1093. 163. Coiera E. Four rules for the reinvention of health care. BMJ 2004;328:1197- 1199. 164. Fieschi M, Dufour JC, Staccini P, Gouvernet J, Bouhaddou O. Medical decision support systems: old dilemmas and new paradigms? Methods Inf Med 2003;42:190-198. 165. Sintchenko V, Magrabi F, Tipper S. Are we measuring the right end-points? Variables that affect the impact of computerised decision support on patient outcomes: a systematic review. Med Inform Internet Med 2007;32:225-240. 166. Forrest GN, Van Schooneveld TC, Kullar R et al. Use of electronic health records and clinical decision support systems for antimicrobial stewardship. Clin Infect Dis 2014;59 Suppl 3:S122-133. 167. Drekonja DM, Filice GA, Greer N et al. Antimicrobial stewardship in outpatient settings: a systematic review. Infect Control Hosp Epidemiol 2015;36:142-152.

BACK TO TABLE OF CONTENTS

37 BACK TO TABLE OF CONTENTS

Chapter 2:

USING BIG DATA TO INTERROGATE ANTIMICROBIAL USE IN COMPANION ANIMALS IN THE COMMUNITY

3

8 Population wide assessment of antimicrobial use in companion animals using a novel data source – a cohort study using pet insurance data LY Hardefeldt*a,b, J Selingerc, MA Stevensona, JR Gilkersona, H Crabba,b, H Billman- Jacobea,b, K Thurskyb, KE Baileya,b M Awadc and GF Browninga,b a Asia-Pacific Centre for Animal Health, Melbourne Veterinary School, Department of Veterinary Biosciences, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria, Australia b National Centre for Antimicrobial Stewardship, Peter Doherty Institute, Grattan St, Carlton, Victoria, Australia, and Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria, Australia c PetSure (Australia) Pty Ltd, Chatswood, NSW, Australia

This paper is under consideration for publication.

Abstract Background Antimicrobial use in veterinary practice is under increasing scrutiny as a contributor to the rising risk of multidrug resistant bacterial pathogens. Surveillance of antimicrobial use in food animals is extensive, but population level data is lacking for companion animals. Lack of census data means cohorts are usually restricted to those attending veterinary practices, which precludes aggregating data from large cohorts of animals, independent of their need for veterinary intervention. Methods A retrospective cohort study was performed using a novel data source; a pet insurance database. The rate of antimicrobial prescribing, and the rate of prescribing of critically important antimicrobials, was measured in a large population of dogs (813,172 dog-years) and cats (129,232 cat-years) from 2013 - 2017. Findings The incidence rate of antimicrobial prescribing was 5·8 prescriptions per 10 dog years (95% CI 5·8-5·9 per 10 dog years) and 3·1 prescriptions per 10 cat years (95% CI 3·1-3·2 per 10 cat years). Critically important antimicrobials accounted for 8% of all the antimicrobials prescribed over the 4-year study. Cats were 4·8- fold more likely than dogs to be prescribed 3rd-generation cephalosporins. Interpretation The level of antimicrobial exposure in dogs and cats was less than half that for the coincident human community. Data such as th provides a unique opportunity to monitor antimicrobial prescribing in veterinary medicine, which is a critical component of optimal antimicrobial stewardship. ese

Introduction Companion animals are often in close contact with humans and can be a source or a recipient of antimicrobial resistant bacteria.1, 2,3, 4,5 In addition, resistant bacterial pathogens have become an increasing problem in veterinary practice, with the emergence of methicillin resistant Staphylococcus pseudintermedius being a particular risk in dogs and cats.6,7 Other resistant pathogens of importance in human medicine have also been detected in companion animals.8,9

39 Surveys of veterinary practitioners have suggested that overuse of antimicrobials after routine surgical procedures10 and for treatment of some medical conditions11-14 occurs. The reported use of critically important antimicrobials (CIA)15 in companion animal veterinary practice varies globally.10 13 CIAs reported most commonly in companion animal practice are fluoroquinolones, predominantly enrofloxacin, and 3rd generation cephalosporins, predominantly cefovecin.11-13 However surveys have mainly relied on self-reporting of prescribing intentions by practitioners when presented with hypothetical scenarios. Objective assessment of antimicrobial use, performed across large populations of companion animals, is lacking. Studies on usage based on assessments of clinical records have been conducted from a relatively limited number of companion animal practices.13,16 In this study we used a novel data source to perform the first comprehensive objective study of antimicrobial use across a large population of dogs and cats, and thus for the first time were able to accurately assess the level of exposure to antimicrobials in a population of dogs and cats.

Materials and methods Data source We conducted a retrospective analysis of pet insurance files from 2013 to 2017. The data files were compiled by data analysis technicians from claim information provided by veterinarians. Veterinarians submitted standardised claims that included clinical information that was most commonly extracted from clinical records by the veterinarians themselves, or the data was sent automatically to the insurer if practices had compliant practice-management software. Data in the clinical notes was sent to the insurance company but was not accessed for this project.

Cohort We identified all dogs and cats that were insured between 2013 and 2017. Exposure was defined as the presence of clinical history with an invoice indicating that the animal had visited a veterinarian and had a claim submitted. The outcomes of interest were administration or prescription of a systemic antimicrobial, and of a systemic antimicrobial with a high-importance rating. The diagnosis and the use of diagnostic pathology testing was also recorded. Confounding variables that were evaluated included species, location (rural or metropolitan), and state. Antimicrobials applied topically were excluded from further analysis.

Statistics Descriptive statistics were computed, with percentages reported as a proportion of animals in the data set that received antimicrobial therapy. Differences in proportions were tested using a Chi squared test. A logistic regression model was used to identify factors that were associated with antimicrobial therapy. The explanatory variables assessed in the model included location (rural or metropolitan, and state of Australia), species and time. The outcome of interest was a proportion, where the numerator was the count of antimicrobial therapies and the denominator was either the total number of claims submitted or the total number of pets insured. A Poisson regression model was used to identify factors

40 that were associated with the incidence rate of antimicrobial therapy in the insured population (exposure to antimicrobials). The explanatory variables assessed in the model included species, month and year.

Unconditional associations between each of the hypothesised explanatory variables and the outcome of interest were computed using an odds ratio or an incidence rate ratio. Explanatory variables with unconditional associations significant at P < 0·20 (2-sided) were selected for multivariable modelling. For the multivariable model the outcome of interest was parameterised as a function of the explanatory variables with unconditional associations significant at P < 0·20, as described above. Explanatory variables that were not significant were then removed from the model one at a time, beginning with the least significant, until the estimated regression coefficients for all explanatory variables retained were significant at an alpha level of less than 0·05. Explanatory variables that were excluded at the initial screening stage were tested for inclusion in the final model and were retained in the model if their inclusion changed any of the estimated regression coefficients by more than 20%.

Biologically plausible two-way interactions were tested and none were significant at an alpha level of 0·05. Generalised logistic regression and Poisson regression models were fitted, and Chi squared tests performed, using functions within Stata v13.

Results There were 222,069 dogs and 37,732 cats registered in the database in 2013. This increased to 385,915 dogs and 60,807 cats over the study period to the end of 2016, which equated to 813,172 dog-years and 129,232 cat-years studied. There were estimated to be 4·8 million pet dogs and 3·8 million pet cats in Australia in 201617, suggesting that the insurance database studied contained 8% of the dog population and 1·6% of the cat population.

In total, 1,919,382 insurance claims were made over the study period, with 531,018 including the prescription of one or more antimicrobial agents. A total of 611,788 courses of antimicrobial treatment were prescribed. The average proportion of animals exposed to antimicrobials was 187 per 1000 dogs and 102 per 1000 cats each year over the 4-year period. The incidence rate of exposure to antimicrobials was 5.8 prescriptions per 10 dog years (95% CI 5·8-5·9 per 10 dog years) and 3·1 prescriptions per 10 cat years (95% CI 3·1-3·2 per 10 cat years). Cats had a 47% lower rate of exposure to antimicrobials than dogs (RR 0·53, 95% CI 0·53-0·54, P<0·001).

Claims were submitted on average for 35% of insured dogs and 21% of insured cats each year. Cats were 49% less likely than dogs to have a claim submitted (OR 0·51, 95% CI 0·50-0·51, P<0.001). The odds of an animal having a claim submitted increased by 2·3% each year (OR 1·02, 95% CI 1·02-1·03, P<0·001). The odds of an animal in a metropolitan area having a claim submitted were 35% higher than for those in rural areas (OR 1·35, 95% CI 1·34-1·37, P<0·001).

41 Among animals that had an insurance claim submitted, other than for routine preventative health measures (vaccination, parasite control, desexing), 53% received systemic antimicrobials (48% of cats and 54% of dogs). After adjusting for location and year, there was no difference in the odds of cats and dogs, with a claim submitted, being administered antimicrobials (OR 0·99, 95% CI 0·97-1·01, P=0·179) (Table 1). However, we found that cats were 5.7-fold more likely to be administered a critically important antimicrobial (OR 5·7, 95% CI 5·5-5·8, P<0·001), most commonly cefovecin (Table 2).

42 Routine preventative health measures (vaccination, parasite control, desexing) accounted for 22% of the claims for dogs and 27% of the claims for cats. Of the claims not associated with preventative health, non-infectious orthopaedic disorders (14%), dermatitis (5·9%) and ear disease (5·8 %) were the most common reasons for claims in dogs, while wounds (6·8%), lower urinary tract disease (4·7%) and abscesses (3·8%) were the most common reasons for claims for cats. Dermatitis, wounds, and non-infectious orthopaedic disorders were the most common reasons antimicrobials were prescribed to dogs (12%, 8·1%, and 6·3% of claims including prescription of an antimicrobial respectively), while wounds, abscesses, and lower urinary tract disease were most common in cats (16%, 9·8%, and 6·3% of claims including prescription of an antimicrobial respectively). In dogs, amoxycillin/clavulanate (34%), cephalexin (19%) and metronidazole (10%) were the most frequently prescribed antimicrobials (Figure 1), while in cats amoxycillin/clavulanate (33%), cefovecin (29%) and doxycycline (8%) were the most frequently prescribed antimicrobials (Figure 2). Narrow spectrum antimicrobials accounted for 35% of the antimicrobials used in dogs and 18% of the antimicrobials used in cats.

35000

30000

25000

20000

15000 No. prescriptionsNo. 10000

5000

0

Amoxycillin/Clavulanate Cephalexin Metronidazole Cefovecin Amoxycillin Clindamycin Doxycycline Enrofloxacin Penicillin Fig. 1. Number of antimicrobial prescriptions for the most common conditions affecting dogs between 2013 and 2017.

4

3 3000

2500

2000

1500

No. prescriptionsNo. 1000

500

0

Amoxycillin/Clavulanate Cephalexin Metronidazole Cefovecin Amoxycillin Clindamycin Doxycycline Enrofloxacin Penicillin

Fig. 2. Number of antimicrobial prescriptions for the most common conditions affecting cats between 2013 and 2017.

There were 49,052 courses of CIAs prescribed over the study period (8·0% of all antimicrobials prescribed). Cefovecin (56%) and enrofloxacin (40%) were the most frequently prescribed CIAs. With the exception of cefovecin in cats, no other CIA represented more than 5% of the antimicrobial use in a species (Figure 3). The rate of cefovecin use over each of the 4 years of the study was did not vary (RR 0·99, 95% CI 0·98-1·00, P=0·065). Cefovecin was used most frequently for wounds and abscesses in cats (22% and 13% of cefovecin prescriptions, respectively), both of which can be managed without antimicrobials in uncomplicated cases. In dogs, cefovecin was most commonly used for dermatitis and dental conditions (20% and 7% cefovecin prescriptions, respectively). Overall, 6·2% of the cats and 8·2% of dogs that received cefovecin had pathology of any kind performed in any consultation linked to this treatment.

Cats had a 54% lower odds of prescription of fluoroquinolone than dogs (OR 0·46, 95% CI 0·43-0·48, P<0·001). There was a small, but significant, increase in the rate of fluoroquinolone use over the four years of the study, after adjusting for species (OR 1·02, 95% CI 1·003-1·03, P=0·014). Enrofloxacin was prescribed most frequently for ear disease, non-infectious orthopaedic disorders and gastroenteritis in dogs (8·2%, 4·2% and 2·9% of enrofloxacin prescriptions, respectively) and for wounds, trauma and urinary tract disease in cats (6·6%, 4·4% and 3·9% of enrofloxacin prescriptions, respectively).

4

4 35

30

25

20

15

10

5

Proportion of total antimicrobials (%) 0 Cats Dogs Cats Dogs Cats Dogs Cats Dogs 2013 2014 2015 2016 Cefovecin Enrofloxacin Other fluoroquinolones

Fig. 3. Proportion of use of critically important antimicrobials over time for dogs and cats.

A seasonal influence on prescribing was seen for dogs (Figure 4) and cats (Figure 5). The rate of antimicrobial prescribing in dogs was 13% higher in spring (RR 1·13, 95% CI 1·12-1·14, P<0·001) and 12% higher in summer (RR 1·12, 95% CI 1·11-1·13, P<0.001) than in winter. The rate of antimicrobial prescribing in cats was 6% lower in summer (RR 0·94, 95% CI 0·92 - 0·97, P<0·001) than in winter. Dogs and cats that had a claim submitted had a 6·4% higher odds of receiving an antimicrobial if they were presented to a veterinarian in a metropolitan area than in a rural area (RR 1·06, 95% CI 1·05-1·08, P<0·001) (Table S1). There was a small, but significant, reduction in the year-on-year rate of exposure to antimicrobials, after adjusting for species (RR 0·99, 95% CI 0·986-0·997, P=0·002).

8

7

6

5

4

3 (per 10 animal years) animal10 (per 2

Incidence of antimicrobial prescriptions antimicrobialof Incidence 1

0 Jul Jul Jul Jul Jan Jan Jan Jan Sep Sep Sep Sep Mar Mar Mar Mar Nov Nov Nov Nov May May May May 2013 2014 2015 2016

Amoxycillin/Clavulanic Acid Cephalexin Metronidazole

Enrofloxacin Amoxycillin Doxycycline

Cefovecin Total antimicrobials Fig. 4. Monthly incidence rate of antimicrobial prescribing in dogs between 2013 to 2017.

4

5 4.5 4 3.5

prescriptions 3 2.5 2

(per 10 cat years)cat10 (per 1.5 1 0.5 Incidence of antimicrobialof Incidence 0 Jul Jul Jul Jul Jan Jan Jan Jan Sep Sep Sep Sep Mar Mar Mar Mar Nov Nov Nov Nov May May May May 2013 2014 2015 2016

Amoxycillin/Clavulanic Acid Cephalexin Metronidazole

Enrofloxacin Amoxycillin Doxycycline

Cefovecin Total antimicrobials Fig. 5. Monthly incidence rate of antimicrobial prescribing in cats between 2013 and 2017

Discussion Data such as this provides a unique opportunity to monitor antimicrobial prescribing in veterinary medicine, which is a critical component of optimal antimicrobial stewardship. This study uses a novel data source to deliver the most comprehensive objective report to date of antimicrobial use in dogs and cats. In addition, this data is the first to report exposure to antimicrobials in a population of dogs and cats. Dogs were more likely to be treated with antimicrobials than cats, with 5·8 prescriptions per 10 dog years and only 3·1 prescriptions per 10 cat years. However, over half of the animals in this cohort that presented to a veterinarian for non-routine examination were treated with antimicrobials. Of particular concern was the observation that more than 25% of all antimicrobials used in cats were the 3rd generation cephalosporin, cefovecin. In contrast, in a study across 11 practices in the United Kingdom (UK), cats were more likely to be treated with antimicrobials13.

However, there was a significant decline in both the rate of total antimicrobial exposure in the population (1% in each year of the study). This decline, while encouraging, is much less than reported over a 2-year period in 457 sentinel companion animal practices in the UK18. Exposure to antimicrobials in this cohort of animals was much lower than community antimicrobial use in humans in Australia, where, in 2014, 1164 prescriptions were issued per 1000 people 19, and in the United States20, where 867 prescriptions are issued per 1000 people per year.

The use of defined daily doses in reporting of antimicrobial prescribing in humans prevents further global comparisons, as these cannot be calculated in veterinary medicine, where patient size differs considerably.

4

6 This is the first report of seasonal differences in the rate of exposure to antimicrobials in veterinary medicine. The pattern in prescribing differs from the typical pattern in community medical prescribing, where peaks in antimicrobial use occur in winter, due to respiratory tract infections19,21-23. The pattern seen in dogs could be attributable to peaks in seasonal diseases, such as allergic dermatitis, seen in warmer months. However, the seasonal peak in prescribing in cats warrants further investigation.

Animals from urban areas had 35% higher odds of having a claim submitted and 6.3% higher odds of having an antimicrobial prescribed compared to animals from rural areas. Although socioeconomic differences might be expected between rural and metropolitan areas, which may influence the likelihood that an animal is presented to a veterinarian, in an insured population these differences are likely negated. Increased antimicrobial prescribing in metropolitan areas may reflect differences in disease processes occurring in metropolitan and rural environments (i.e. most metropolitan areas in Australia are coastal), or may be due to differences in expectations of animal owners or in the attitudes to antimicrobial prescribing among veterinarians.

While antimicrobial use was lower in cats in this cohort compared to dogs, cats had a 5·7-fold higher odds of receiving cefovecin, a trend also identified in the UK13. Cefovecin use in cats was most commonly for wounds and abscesses, despite guidelines recommending either no antimicrobial therapy24,25 or amoxycillin/clavulanate26. The cefovecin label dictates use only when culture and susceptibility indicates that this drug is the only option27, but there was little evidence from the claims that this was followed in most instances. However, cats can be difficult to medicate, and use of cefovecin is likely to be driven by convenience and compliance concerns (a single injection provides effective serum concentrations for 14 days27) as reported previously28.

Critically important antimicrobials accounted for 8% of all the antimicrobials prescribed over the study period. While restricting all off-label use of antimicrobials in animals in Australia is likely to be detrimental to antimicrobial stewardship measures29, and animal welfare in general, it may be necessary to restrict the use of 3rd generation cephalosporins in this manner to reduce the inappropriate use of this antimicrobial. Narrow spectrum antimicrobials accounted for 35% of the antimicrobials used in dogs and 18% of those used in cats, consistent with other veterinary studies13, but considerably greater than use of narrow spectrum agents in community medical practice in Australia (where 8% of prescriptions are for narrow spectrum agents)19. The preferential use of narrow spectrum antimicrobials is recommended in many antimicrobial stewardship programs, including a program recently designed for veterinary practices in Australia30. In dogs and cats the most common conditions for which amoxycillin/clavulanate was prescribed was for wounds, despite guidelines recommending either no antimicrobial therapy24,25 or amoxycillin alone26.

As there is less likely to be a financial disincentive to seeking veterinary attention, a higher proportion of animals in an insured population might be expected to be

4

7 presented to a veterinarian, and the proportion of consultations that result in an antimicrobial prescription may also be higher. However, the patterns of antimicrobial prescribing are likely to closely reflect those of the greater population and thus the data can be assumed to provide a reliable upper estimate of antimicrobial use across the remainder of the population. The estimates of the proportion of dogs and cats treated with an antimicrobial following a veterinary consultation in our study were similar to those seen in smaller studies in Canada12 and similar to those seen in cats, but higher than for dogs, in the UK13.

In conclusion, this study has demonstrated the value of insurance claims as a novel data source for estimating population wide exposure to antimicrobials among companion animals. This objective data provides a comprehensive overview of antimicrobial use in dogs and cats and highlights areas where education is needed, and further investigation is warranted. Furthermore, these data will assist in the implementation and monitoring of antimicrobial stewardship programs. The level of off-label use of 3rd generation cephalosporins detected in cats in our study is of particular concern, suggesting that consideration needs to be given to strictly restricting the use of cefovecin to its labelled indications.

Acknowledgements We thank PetSure for supplying the data.

References 1. Guardabassi L, Schwarz S, Lloyd DH. Pet animals as reservoirs of antimicrobial- resistant bacteria. J Antimicrob Chemother 2004; 54(2): 321-32. 2. Pantosti A. Methicillin-Resistant Staphylococcus aureus associated with animals and its relevance to human health. Front Microbiol 2012; 3: 127. 3. Couto N, Monchique C, Belas A, Marques C, Gama LT, Pomba C. Trends and molecular mechanisms of antimicrobial resistance in clinical Staphylococci isolated from companion animals over a 16 year period. J Antimicrob Chemother 2016; 71(6): 1479-87. 4. Schwaber MJ, Navon-Venezia S, Masarwa S, et al. Clonal transmission of a rare methicillin-resistant Staphylococcus aureus genotype between horses and staff at a veterinary teaching hospital. Vet Microbiol 2013; 162(2-4): 907-11. 5. Paul NC, Moodley A, Ghibaudo G, Guardabassi L. Carriage of methicillin-resistant Staphylococcus pseudintermedius in small animal veterinarians: indirect evidence of zoonotic transmission. Zoonoses Public Health 2011; 58(8): 533-9. 6. Beck KM, Waisglass SE, Dick HL, Weese JS. Prevalence of meticillin-resistant Staphylococcus pseudintermedius (MRSP) from skin and carriage sites of dogs after treatment of their meticillin-resistant or meticillin-sensitive Staphylococcal pyoderma. Vet Dermatol 2012; 23(4): 369-75, e66-7. 7. Saputra S, Jordan D, Worthing KA, et al. Antimicrobial resistance in coagulase- positive Staphylococci isolated from companion animals in Australia: A one year study. PLoS One 2017; 12(4): e0176379. 8. Platell JL, Cobbold RN, Johnson JR, et al. Commonality among fluoroquinolone- resistant sequence type ST131 extraintestinal Escherichia coli isolates from humans and companion animals in Australia. Antimicrob Agents Chemother 2011; 55(8): 3782-7. 9. Abraham S, O'Dea M, Trott DJ, et al. Isolation and plasmid characterization of carbapenemase (IMP-4) producing Salmonella enterica Typhimurium from cats. Sci Rep 2016; 6: 35527.

4

8 Hardefeldt LY, Browning GF, Thursky K, et al. Antimicrobials used for surgical prophylaxis by companion animal veterinarians in Australia. Veterinary 10. Microbiology 2017; 203: 301-7. Hardefeldt LY, Holloway S, Trott DJ, et al. Antimicrobial prescribing in dogs and cats in Australia: results of the Australasian Infectious Disease Advisory Panel survey. J 11. Vet Intern Med 2017; 31(4): 1100-7. Murphy CP, Reid-Smith RJ, Boerlin P, et al. Out-patient antimicrobial drug use in dogs and cats for new disease events from community companion animal practices 12. in Ontario. Canadian Veterinary Journal-Revue Veterinaire Canadienne 2012; 53(3): 8. Mateus A, Brodbelt DC, Barber N, Stark KD. Antimicrobial usage in dogs and cats in 291-first opinion veterinary practices in the UK. J Small Anim Pract 2011; 52(10): 515-21. 13. Pleydell EJ, Souphavanh K, Hill KE, French NP, Prattley DJ. Descriptive epidemiological study of the use of antimicrobial drugs by companion animal 14. veterinarians in New Zealand. N Z Vet J 2012; 60(2): 115-22. Australian Strategic and Technical Advisory Group on Antimicrobial Resistance. Importance rating and summary of antibacterials used in human health in Australia. 15. http://www.health.gov.au/internet/main/publishing.nsf/Content/ohp-amr.htm : Commonweath of Australia, 2015. Radford AD, Noble PJ, Coyne KP, et al. Antibacterial prescribing patterns in small animal veterinary practice identified via SAVSNET: the small animal veterinary 16. surveillance network. Vet Rec 2011; 169(12): 310. Animal Medicines Australia. Pet ownership in Australia. http://animalmedicinesaustralia.org.au/wp-content/uploads/2016/11/AMA_Pet- 17. Ownership-in-Australia-2016-Report_sml.pdf, 2016. Singleton DA, Sanchez-Vizcaino F, Dawson S, et al. Patterns of antimicrobial agent prescription in a sentinel population of canine and feline veterinary practices in the 18. United Kingdom. Vet J 2017; 224: 18-24. Australian Commission on Safety and Quality in Health Care. AURA 2016: The first report on antimicrobial use and resistance in human health. Sydney: ACSQHC, 2016. 19. Suda KJ, Hicks LA, Roberts RM, Hunkler RJ, Taylor TH. Trends and seasonal variation in outpatient antibiotic prescription rates in the United States, 2006 to 2010. 20. Antimicrob Agents Chemother 2014; 58(5): 2763-6. Williamson DA, Roos R, Verrall A, Smith A, Thomas MG. Trends, demographics and disparities in outpatient antibiotic consumption in New Zealand: a national study. J 21. Antimicrob Chemother 2016; 71(12): 3593-8. Goossens H, Ferech M, Vander Stichele R, Elseviers M, Group EP. Outpatient antibiotic use in Europe and association with resistance: a cross-national database 22. study. Lancet 2005; 365(9459): 579-87. Sun L, Klein EY, Laxminarayan R. Seasonality and temporal correlation between community antibiotic use and resistance in the United States. Clin Infect Dis 2012; 23. 55(5): 687-94. Asia Pacific Centre for Animal Health, National Centre for Antimicrobial Stewardship. Australian Veterinary Prescribing Guidelines. 2017. 24. www.fvas.unimelb.edu.au/vetantibiotics (accessed 13/9/17). Spohr A, Schjoth B, Wiinberg B, et al. Antibiotic Use Guidelines for Companion Animal Practice. www.fecava.org/sites/default/files/DSAVA_AntibioticGuidelines 25. v1-1_3(1).pdf: Danish Small Animal Veterinary Association, 2009. British Small Animal Veterinary Association. PROTECT. 2016. . 26. Zoetis Australia Pty Ltd. Convenia. 2013. http://websvr.infopest.com.au/LabelRouter?LabelType=L&Mode=1&Produhttps://www.bsava.com/Resources/Veterinary-resources/PROTECT ctCode= 27. 60461 (accessed 2/10/17).

4

9 28. Mateus AL, Brodbelt DC, Barber N, Stark KD. Qualitative study of factors associated with antimicrobial usage in seven small animal veterinary practices in the UK. Prev Vet Med 2014; 117(1): 68-78. 29. Hardefeldt LY, Gilkerson JR, Billman-Jacobe H, et al. Antimicrobial labelling in Australia: a threat to antimicrobial stewardship. Aust Vet J 2017; accepted for publication. 30. Asia Pacific Centre for Animal Health, National Centre for Antimicrobial Stewardship. Veterinary Antimicrobial Stewardship Program procedure. 2017. www.fvas.unimelb.edu.au/vetantibiotics//about/antimicrobial-stewardship (accessed 5/10/17).

50 BACK TO TABLE OF CONTENTS

Chapter 3:

ASSESSMENT OF HISTORIC ANTIMICROBIAL PRESCRIBING PATTERNS BY VETERINARIANS IN COMPANION ANIMAL PRACTICE

51 Standard Article J Vet Intern Med 2017;31:1100–1107

Antimicrobial Prescribing in Dogs and Cats in Australia: Results of the Australasian Infectious Disease Advisory Panel Survey

L.Y. Hardefeldt , S. Holloway, D.J. Trott, M. Shipstone, V.R. Barrs, R. Malik, M. Burrows, S. Armstrong, G.F. Browning, and M. Stevenson

Background: Investigations of antimicrobial use in companion animals are limited. With the growing recognition of the need for improved antimicrobial stewardship, there is urgent need for more detailed understanding of the patterns of antimi- crobial use in this sector. Objectives: To investigate antimicrobial use for medical and surgical conditions in dogs and cats by Australian veterinarians. Methods: A cross-sectional study was performed over 4 months in 2011. Respondents were asked about their choices of antimicrobials for empirical therapy of diseases in dogs and cats, duration of therapy, and selection based on culture and sus- ceptibility testing, for common conditions framed as case scenarios: 11 medical, 2 surgical, and 8 dermatological. Results: A total of 892 of the 1,029 members of the Australian veterinary profession that completed the survey satisfied the selection criteria. Empirical antimicrobial therapy was more common for acute conditions (76%) than chronic conditions (24%). Overall, the most common antimicrobial classes were potentiated aminopenicillins (36%), fluoroquinolones (15%), first- and second-generation cephalosporins (14%), and tetracyclines (11%). Third-generation cephalosporins were more fre- quently used in cats (16%) compared to dogs (2%). Agreement with Australasian Infectious Disease Advisory Panel (AIDAP) guidelines (generated subsequently) was variable ranging from 0 to 69% between conditions. Conclusions and Clinical Importance: Choice of antimicrobials by Australian veterinary practitioners was generally appro- priate, with relatively low use of drugs of high importance, except for the empirical use of fluoroquinolones in dogs, particu- larly for otitis externa and 3rd-generation cephalosporins in cats. Future surveys will determine whether introduction of the 2013 AIDAP therapeutic guidelines has influenced prescribing habits. Key words: Antibiotic; Companion animals; Stewardship.

ntimicrobial resistance develops in response to Aantimicrobial use1–3 regardless of the animal species Abbreviations: being treated, with greater use likely to contribute to AIDAP Australasian Infectious Disease Advisory Panel development of resistance to multiple drug classes. This is C & S culture and sensitivity a growing threat in human hospitals, the community and IQR interquartile range in companion and production animals. Veterinary IRSAD index for relative socioeconomic advantage- antimicrobial usage, has come under increasing scrutiny disadvantage by medical, public health, and government officials, LUTI lower urinary tract infection

From the Faculty of Veterinary and Agricultural Sciences, Asia- especially in food-producing animals. In companion ani- Pacific Centre for Animal Health, University of Melbourne, mals, an apparent increase, or increased reporting of mul- Melbourne, (Hardefeldt, Browning, Stevenson); Advanced Vetcare, tidrug-resistant pathogens, especially coagulase-positive Kensington, Vic. (Holloway); School of Animal and Veterinary staphylococcal species,4–7 suggests that investigation of Sciences, University of Adelaide, Adelaide, SA (Trott); School of patterns of antimicrobial usage in companion animal Veterinary Sciences, University of Queensland, Gatton, Qld practice is needed. Since the registration of fluoro- (Shipstone); Faculty of Veterinary Science, University of Sydney, Sydney, NSW (Barrs, Malik); Animal Dermatology, Perth, WA quinolones (ie, enrofloxacin, marbofloxacin, difloxacin, (Burrows); and the Zoetis Animal Health, Rhodes, NSW Australia orbifloxacin, and most recently pradofloxacin) starting in (Armstrong). 1989 and an injectable long-acting 3rd-generation cepha- The survey was designed by the AIDAP panel and undertaken by losporin (cefovecin) in 2008, for specific use in dogs and Pfizer Animal Health Australia in 2011 (now Zoetis Animal Health cats, antimicrobial usage patterns in Australian compan- Australia). Analysis of the survey data was performed at the Univer- ion animal practice have not been examined. sity of Melbourne. The article has not been presented at any meetings. Data on antimicrobial use in companion animal prac- Corresponding author: L. Hardefeldt, Melbourne Veterinary tice in Australia are limited to a single cross-sectional 8 School, University of Melbourne, Parkville, Vic. 3052, Australia; study carried out in 1997. In this survey, respondents e-mail: [email protected]. were asked about patterns of use of various systemic Submitted December 11, 2016; Revised March 15, 2017; antibacterial drugs and their approach to treatment of 9 Accepted April 11, 2017. specific medical scenarios. Penicillins and cephalospor- Copyright © 2017 The Authors. Journal of Veterinary Internal ins were the most commonly used drugs, with amoxi- Medicine published by Wiley Periodicals, Inc. on behalf of the Ameri- can College of Veterinary Internal Medicine. cillin-clavulanate the most frequently prescribed This is an open access article under the terms of the Creative antimicrobial agent. Empiric antibiotic therapy was Commons Attribution-NonCommercial License, which permits use, used in the vast majority of acute medical conditions distribution and reproduction in any medium, provided the original (76–94% of cases) and was frequently used in chronic work is properly cited and is not used for commercial purposes. conditions (15–50% of cases). DOI: 10.1111/jvim.14733

52 Antimicrobial Prescribing by Australian Veterinarians 1101

The Australian Strategic and Technical Advisory 11 specific medical disorders when clinical evidence sug- Group on Antimicrobial Resistance have issued an gested the presumptive diagnosis. They were also asked importance rating and summary of antibacterials used about their approaches to 2 surgical conditions; routine in human health in Australia in 2015. Those given a desexing and dental scaling and polishing, with tooth high importance rating include piperacillin-tazobactam, extractions. The specified disorders included abscess/cel- ticarcillin-clavulanate (now no longer manufactured but lulitis, chronic gingivostomatitis/“faucitis,” acute febrile available at the time of the survey), the 3rd- and 4th- illness, peritonitis, chronic rhinosinusitis, pyothorax, generation cephalosporins, aztreonam, tigecycline, van- acute upper and lower respiratory tract infections, acute comycin, teicoplanin, amikacin, the streptogramins (eg, and recurrent lower urinary tract infections (LUTI)/cys- pristinamycin), fluoroquinolones, and rifampicin.9 These titis and LUTI with concurrent chronic kidney disease. antimicrobials should be treated as third-line therapies The third section asked about management of selected and should only be used where culture and susceptibil- dermatological conditions and otitis externa, including ity (C & S) testing or other compelling clinical evidence surface, superficial and deep pyodermas, dermatophyto- indicates their use. Of the antimicrobials with a high sis, and superficial yeast infections of the skin, as well importance rating, only the 3rd-generation cephalospor- as uncomplicated and refractory otitis externa. Both ins and fluoroquinolones are registered for use in dogs open and closed questions were used. Drop-down and cats in Australia. menus provided lists of commercially available antimi- The Australasian Infectious Disease Advisory crobials from which respondents could select their Panel (AIDAP) was convened with a view to developing favored therapy. antimicrobial and therapeutic guidelines for common Data were downloaded from the Website to spread- medical, surgical and dermatological conditions seen in sheets (Microsoft Office Access, Microsoft Office Excel). general veterinary practice in Australia. These guidelines Any questions not completed by a respondent were were released in 2013 and include evidence-based rec- excluded from the analysis of that question. Simple ommendations, where possible, and specialist veterinary descriptive statistics were computed with percentages opinion where there was a limited evidence base. being reported as a proportion of the total number of The aims of this study were to investigate empirical respondents answering a particular question. Given that antimicrobial use (ie, drug choices), the frequency of this study did not use a simple random sampling design, use of C & S testing as a tool for selecting antimicro- data were analyzed to account for overrepresentation bials, and the proportion of “high importance” rating by state of practice.10 Sampling weights provided an antimicrobials, through case scenario presentations to estimate of the inverse probability of a veterinarian’s identify likely practitioner prescribing behavior. A sec- involvement in the survey, WHi, and were quantified as ondary aim was to determine the frequency of agree- follows: ment of antimicrobial use with the AIDAP therapeutic guidelines which were generated after the survey had Ni WHi been conducted. ¼ ni;

Methods where Ni is the number of registered veterinarians in The source population for the survey was clinicians the state in 2011, and n is the number of veterinarians practicing veterinary medicine in Australia in 2011. At from that state who completed the survey. Throughout that time, there were an estimated 7,300 registered vet- this article, all profession level data are described using erinarians in Australia. To be 95% certain that this esti- adjusted values based on survey design, sampling mate of the prevalence of veterinarians using a given weights, and finite correction factors. Proportions of class of antimicrobial was within 5% of a true preva- questionnaire responses are reported as unadjusted lence of 50%, a total of 365 completed surveys were counts. required. Sample size calculations were carried out The index for relative socioeconomic advantage-dis- assuming a 50% prevalence because this provided the advantage (IRSAD) and usual resident population of largest sample size estimate for a constant margin of each postcode for participants in the survey was error. Respondents were self-selected and were encour- accessed from the Australian Bureau of Statistics.11 aged to participate through a variety of electronic and Regression models were used to quantify the associa- print media sources over a 4-month period in 2011. tion between individual respondent-level variables (year The survey was created online by web-based designers of graduation, percentage of small versus large animal in coordination with the AIDAP (questionnaire avail- practice, IRSAD) and the probability of a veterinarian able as Supporting Information). There were 3 sections. prescribing in such a way that agreed with the AIDAP The first section asked for the respondent’s veterinary guidelines. For continuously distributed explanatory board registration number, year of graduation, and an variables, a Shapiro-Wilk test for normality was used. estimate of the proportion of clinical work they per- Differences in independent medians were assessed using formed on cats, dogs, horses, production animal, or the Mann-Whitney tests. A binary logistic regression other species. In the second section, respondents were model was developed with year of graduation expressed asked to indicate their usual (>50% of the time) as a 2-level categorical variable: <5 years since gradua- approach to the treatment of cats and dogs for each of tion and 5 or greater years since graduation. The

53 1102 Hardefeldt et al proportion of time spent on small animal practice was dogs (95% CI 22–28%). There was no difference in the expressed as a 2-level categorical variable: those that choice of antimicrobial therapy between dogs and cats, spent more than 70% of their time working with com- with more than 90% of respondents indicating the use panion animals (“companion animal practitioners”) and of aminopenicillins (51%), other b-lactam drugs (25%) those that spent 70% or less of their time working or potentiated aminopenicillins (17%). High importance with companion animals (“mixed animal practitioners”). rated antimicrobials were used by only 11 respondents The binary outcome variable for this analysis was for this indication (3.6%); 6 used 3rd-generation cepha- whether or not the reported antimicrobial usage pat- losporins (4 in cats, 2 in dogs), 3 reported using ticar- terns reported by the respondent were consistent with cillin-clavulanate (1 in cats, 2 in dogs), and 2 reported AIDAP guidelines or not. Descriptive analysis and the using enrofloxacin (1 in cats, 1 in dogs). Duration of logistic regression analysis were carried out using Stata therapy did not differ between dogs and cats with a version 13a . median duration of therapy of 2 days (Q1–Q3 1– This study was organized and sponsored by a veteri- 3 days). nary pharmaceutical company, and ethics clearance was Most respondents used antimicrobials for dental pro- not required by the University of Melbourne as no cedures, with extractions, in both cats (95% [639 of identifying information was used in the analysis. 675], 95% CI 93–96%) and dogs (94% [605 of 643], 95% CI 92–96%), and selection was empiric in 94% of Results cat cases (95% CI 92–96%) and 94% of dog cases (95% CI 92–96%). Antimicrobial therapy was initiated A total of 1,029 Australian veterinary practitioners before dentistry by 64% (665 of 1,041, 95% CI 61– completed the survey. Of these, 892 satisfied the selec- 67%) of practitioners, and the median duration of ther- tion criteria for inclusion, representing more than 12% apy was 7 days (Q1–Q3 7–10 days). The choice of of the total number of registered veterinarians in 2011. antimicrobial differed between dogs and cats with All states and territories were represented, as were potentiated aminopenicillins (33%), clindamycin (30%), recent and older graduates. More than 70% of respon- and 3rd-generation cephalosporins (21%) used most fre- dents were companion animal practitioners. Only 4.5% quently in cats, and potentiated aminopenicillins (46%) of respondents had a caseload in which <50% of and clindamycin (35%) used most frequently in dogs. patients were dogs and cats. High importance rating antimicrobials were used by As there was no difference in the frequency with 13% of respondents, with the vast majority being a 3rd- which empirical antimicrobial therapy was used rather generation cephalosporin (190 of 203, 94%) which were than antimicrobial therapy directed by the results of C predominately administered to cats (168 of 190, 88%). & S testing between dogs and cats, the results were Overall there were 22,748 antimicrobial therapies combined for each question. Overall, antimicrobial reported across the scenarios. The most commonly used selections were empirical in 51% (7,290 of 14,414) of antimicrobials were aminopenicillins (41% of dog thera- cases (range 8–79%), guided by C & S in 26% (3,694 of pies and 41% of cat therapies), followed by fluoro- 14,414) of cases (range 0.1–70%), and empirical therapy quinolones (18% of dog therapies and 11% of cat was employed pending the results of C & S testing in therapies), 1st- or 2nd-generation cephalosporins (22% 24% (3,430 of 14,414) (range 13–32%) of cases. There of dog therapies and 3% of cat therapies), and tetracy- were 3 conditions in which C & S testing was used by clines (7% of dog therapies and 17% of cat therapies) >80% of respondents: pyothorax (80%, 95% CI 78– (Table 1). Use of antimicrobials with a high importance 83%), recurrent LUTI (92%, 95% CI 91–94%) and rating ranged from 12 to 47% (median 17%) for cats LUTI with chronic kidney disease (80%, 95% CI 78– and 4 to 42% (median 15%) for dogs among the medi- 82%). Antimicrobial therapy guided by C & S was also cal scenarios. Overall, 3rd-generation cephalosporin use commonly used for chronic rhinosinusitis (67% of was more frequent in cats than dogs (16 versus 1.8%, responses, 95% CI 64–69%). Empirical antimicrobial P < .001) whereas fluoroquinolone use was more fre- therapy was more commonly used for acute conditions quent in dogs (18 versus 11%, P < .001) (Table 1). In (median 63%, quartile 1 [Q1] 55 to quartile 3 [Q3] dogs, fluoroquinolones were also more frequently pre- 77%) than chronic conditions (median 25%, Q1–Q3 scribed for chronic conditions than for acute conditions 17–44%, P = .01). Culture and susceptibility was used (18 and 15% respectively, P < .001). In cats, 3rd-gen- by at least 20% of respondents in all medical scenarios eration cephalosporins were more frequently prescribed including abscesses. There was no difference between for chronic than for acute conditions (18 and 14% mixed and companion animal practitioners in the pro- respectively, P = .001). The amount of fluoroquinolone portion of cases in which C & S was performed (4.7% use was similar in dermatological conditions to medical higher for companion animal practitioners, 95% CI conditions (11%, 95% CI 10–12%), but more frequent 2.5 to 12%, P = .199), although recent graduates in otitis externa (41%, 95% CI 39–43%). In otitis À(<5 years’ experience) used C & S guided antimicrobial externa, where bacterial rods were seen in cytological therapy less commonly than older graduates (6.4% preparations, systemic fluoroquinolone use was reported lower; 95% CI 2.4–11%, P = .002). by 61% (95% CI 58–64%) of respondents. For medical Routine prophylactic antimicrobial therapy was used conditions in dogs, fluoroquinolones were used most by 25% (157 of 631) of respondents for routine desex- frequently to treat pneumonia (29%, 95% CI 27–32%), ing of cats (95% CI 22–28%) and 25% (154 of 618) of pyothorax (31%, 95% CI 27–36%), and recurrent

54 Antimicrobial Prescribing by Australian Veterinarians 1103

Table 1. Overall frequency of antibiotic use across Agreement with AIDAP therapeutic guidelines (post- medical, surgical and dermatological scenarios posed in hoc) was evaluated for use of empirical therapy, use of this survey. antimicrobial therapy guided by C & S or treatment without the use of antimicrobials, as well as drug Frequency (%) Subclass choice, duration of therapy, and overall agreement. Drug Class or Drug Cats Dogs The data were not normally distributed. Overall agree- ment was variable, ranging from 0 to 69% between 1st- and 327 (3.4) 2,908 (22) conditions. There was no difference between medical 2nd-generation cephalosporins and dermatological conditions in the extent of agree- 3rd-generation 1,548 (16) 240 (1.8) ment with the guidelines. The median overall agree- cephalosporins ment was higher for dogs (38%, Q1–Q3 25–46%) than Aminoglycosides Gentamicin 20 (0.2) 54 (0.4) for cats (25%, Q1–Q3 16–31%, P < .001). The overall Amikacin 3 (<0.1) 7 (<0.1) agreement with the guidelines was less than 33% for 4 Total 23 (0.2) 61 (0.5) conditions; gingivostomatitis (0% agreement for cats, b-Lactams Unpotentiated 487 (5.1) 492 (3.7) 6.9% agreement for dogs), pyothorax (3.2% agreement Potentiated 3,330 (35) 4,901 (37) for cats, 0.1% agreement for dogs), peritonitis (0.3% High 46 (0.5) 69 (0.5) agreement for cats, 1.0% agreement for dogs), and importance acute LUTI disease/cystitis (8.8% agreement for cats, rating Total 3,863 (41) 5,462 (41) 16% agreement for dogs). For gingivitis and pyotho- Macrolides 706 (7.4) 617 (4.7) rax, the decision to use empirical antimicrobials or C Chloramphenicol 0 (0) 2 (<0.1) & S testing had much higher agreement with AIDAP Tetracyclines 1,579 (17) 953 (7.2) guidelines, and the poor overall agreement was due to Fluoroquinolones 1,065 (11) 2,389 (18) poor alignment with the recommendations for drug Metronidazole 327 (3.4) 374 (2.8) selection and duration of therapy recommendations Rifampicin 0 (0) 29 (0.2) (Fig 1A,B). For acute cystitis, the poor agreement was Trimethoprim/ 47 (0.5) 161 (1.2) due to the common use of empirical antimicrobial sulfonamides therapy and failure to culture samples from these Other 27 (0.3) 40 (0.3) cases, whereas drug selection and duration of therapy Total 9,512 13,236 were in better agreement (Fig 1C). Finally, for peri- tonitis, there was poor agreement in terms of both empirical choice of drug, use of C & S testing, and LUTI disorders (32%, 95% CI 27–37%). In cats, 3rd- with the selection of drug (Fig 1D), with drugs with a generation cephalosporins were used by more than 25% limited spectrum of activity being chosen for most of respondents in 4 scenarios; cellulitis/abscesses (26%, cases rather than the extended spectrum (usually via 95% CI 24–28%), acute LUTI disease (33%, 95% CI combination therapy) advocated in the AIDAP guideli- 29–36%), recurrent LUTI disease (25%, 95% CI 21– nes (96%, 95% CI 95–97%). Overall, the choice of 31%), and LUTI with concurrent chronic kidney empirical or therapy guided by C & S or treatment disease (26%, 95% CI 22–31%). In contrast, 3rd-gen- without the use of antimicrobials showed the best eration cephalosporins were much less frequently used agreement with the guidelines, with a median of 83% for severe conditions in cats such as pneumonia (7.8%, (Q1–Q3 42–95%). There was no difference between 95% CI 6.3–9.6%), pyothorax (4.2%, 95% CI 2.8– responses about treatment of dogs or cats. The agree- 6.3%), and peritonitis (3.5%, 95% CI 2.4–5.0%). The ment with the guidelines with respect to choice of use of 3rd-generation cephalosporins for dermatological drug, where indicated, did not differ between dogs and cases was rare (1.9% overall, 95% CI 1.5–2.3%). Use cats, with an overall median 43% (Q1–Q3 5–57%). of other antimicrobials with a high importance rating Similarly, agreement with the guidelines on duration of was rare and did not differ between dogs and cats (0.6 therapy and drug selection, where indicated, had the and 0.5%, respectively). The duration of therapy used same level of agreement between dogs and cats (overall by respondents choosing antimicrobials with a high median 36%, Q1–Q3 25–67%). There was no signifi- importance rating did not differ from those choosing cant difference in agreement with AIDAP guidelines antimicrobials with low or medium importance rating. for practitioners who were recent graduates (past The distribution of the number of prescriptions of 5 years) compared to older graduates, nor between antimicrobials of high importance rating for each par- practitioners who were predominately small animal ticipant was positively skewed with lowest 50% of veterinarians compared to veterinarians working in respondents prescribing 12% of these antimicrobials “mixed practices.” and higher 50% prescribing the remaining 88%. The low users of antimicrobials of high importance rating Discussion also used less therapy guided by C & S (38%) than high users (62%, P < .001). There was no difference in popu- This study has shown that empirical antimicrobial lation, IRSAD, year of graduation or percentage time therapy is very common in Australian veterinary in companion animal practice, between low and high practice as is indicated for many conditions both in vet- users of antimicrobials of high importance rating. erinary12 and medical practice.13 This is a similar

5 1104 Hardefeldt et al

B A 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 % Respondants % 20 % Respondants % 20 10 10 0 0 Empiric or C Drug Choice Guideline Empiric or Drug Duration Guideline & S agreed C & S Choice agreed

C D

100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 % Respondants % % Respondants % 20 20 10 10 0 0 Empiric or Drug Duration Guideline Empiric or C Drug Choice Guideline C & S Choice agreed & S agreed

Fig 1. Agreement with Australasian Infectious Disease Advisory Panel guidelines for choice of empirical or antimicrobial therapy guided by culture and susceptibility (C & S), choice of drug and duration of therapy, and overall agreement with the guidelines for treatment of (A) gingivitis, (B) pyothorax, (C) acute cystitis, and (D) peritonitis. White columns indicate treatment choices for cats, and black columns indicate the treatment choices for dogs.

outcome to that reported in a cross-sectional study of improved antimicrobial stewardship and might be practicing veterinarians in New Zealand where therapy, expected to improve clinical outcomes in veterinary guided by C & S, was used in 19% of cases.14 Fluoro- practices, although there are likely to be concerns about quinolones were used empirically at a high rate for its cost-effectiveness for many animal owners. In addi- specific conditions in dogs, as were 3rd-generation tion, this survey did not investigate the methods used cephalosporins in cats. The rate of empiric 3rd-genera- for C & S testing by veterinarians or veterinary labora- tion cephalosporin use in cats in this study is similar to tories. Use of rigorous methodology and veterinary findings by others.15,16 Prophylactic antimicrobial use specific break points is critical for ensuring reliable for routine desexing was less common. The most com- results from C & S testing. Regardless, this promising monly used antimicrobial classes are aminopenicillins, trend may reflect a growing willingness of the public to particularly potentiated aminopenicillins, fluoro- invest in disease investigations, an increase in these ser- quinolones, early-generation cephalosporins, and tetra- vices being offered to clients, increased awareness by cyclines. This is consistent with findings in New the profession of the benefit of testing, and/or an Zealand (amoxicillin-clavulanate 48%, cephalexin increase in treatment failures necessitating further inves- 31%),14 Canada (aminopenicillins 56%, cephalexin tigation. Interestingly, C & S testing was used relatively 33%),16 and the United Kingdom (aminopenicillins frequently in the treatment of abscess (21%). Further 59%, cephalexin 13%).15 Interestingly, there was very investigation is warranted to evaluate the reasoning low use of older broad-spectrum antimicrobials such as behind the high level of C & S testing for this scenario. trimethoprim sulfonamide combinations (0.9%) and The results may indicate a degree of prevarication bias chloramphenicol (0.01%). There was also limited use of in the survey (ie, survey respondents altering their other drugs with a low importance rating, such as answers to survey questions in a way that matches the macrolides (6%), which have traditionally been main- perceived expectations of those carrying out the survey) stays of therapy, particularly in cats. and should be validated. Empirical antimicrobial therapy was less common in The AIDAP therapeutic guidelines recommend the this survey than in the only other survey of antimicro- use of cefovecin, the only 3rd-generation cephalosporin bial usage in dogs and cats in Australia, which was per- registered for use in companion animals in Australia, formed in 1997. In that study, empirical use for acute only for cases where there is likely to be poor compli- conditions ranged from 76 to 94% of cases,8 whereas in ance with oral antimicrobial therapy. As a reflection of this study, the range was 19–79%. The increased use of this, 3rd-generation cephalosporins were much more C & S testing as a tool for directing antimicrobial ther- commonly used in cats compared to dogs by practition- apy that was detected in this survey is likely to reflect ers completing this survey. The most frequent scenarios

6 Antimicrobial Prescribing by Australian Veterinarians 1105 were those in which infection could be effectively trea- economically in their animals, as both C & S testing ted with orally administered antimicrobial agents with a and antimicrobials of high importance rating tend to lower importance rating (ie, cellulitis/bite-wound abscess be expensive in Australian veterinary practices. and LUTI). Further, survey findings highlight the need However, it may also reflect a more proactive clientele for LUTI in cats to be confirmed by in-house micro- that present cases earlier and therefore the need for scopic evaluation of a urine sample before initiating antimicrobials of high importance rating, and directed antimicrobial therapy due to the high prevalence of therapy, is perceived to be less. Further investigation noninfectious cystitis in cats.17–19 The reported high into the factors driving the high use of these antimi- usage of 3rd-generation cephalosporins in cats likely crobials by a selection of the veterinary population is reflects poor compliance in administration of oral drugs warranted. to cats compared to dogs, as cats are less likely to There have been no previous reports on the fre- ingest medications in food, as has been found in a quency of antimicrobial use for routine surgical proce- recent study from the United Kingdom.20 dures in companion animal practice in Australia. Different factors may account for fluoroquinolone Antibiotics are considered unnecessary for routine short administration to dogs. The rate of fluoroquinolone surgeries conducted under sterile conditions, such as administration for some of the scenarios included in this routine desexing.12 Over 75% of respondents in this survey was higher than expected. In complicated canine survey did not use antimicrobial prophylaxis for rou- otitis cases involving Gram-negative pathogens, such as tine desexing. However, with almost one quarter of Pseudomonas aeruginosa with rupture of the tympanic Australian veterinarians still routinely using antimicro- membrane, there are limited therapeutic options and bials for neutering, and the number of these procedures the use of topical fluoroquinolones in this scenario is performed in general practice, this topic requires a often warranted. However, the high frequency of sys- specific education program. Antimicrobials were fre- temic use of fluoroquinolones for both complicated and quently used in patients undergoing dental procedures uncomplicated otitis cases suggests a need for improved including extractions in this survey (90% of respon- antimicrobial stewardship by veterinarians in treating dents). The AIDAP guidelines recommend prophylactic this disease. Awareness by veterinarians of the high antimicrobials if there are extractions or likely to be concentrations of fluoroquinolones that can be achieved bleeding.12 Interestingly, as the release of the AIDAP with topically applied formulations, and hence low risk guidelines, the recommendations for use of antimicro- of resistance development,21 may be lacking. Efforts bials in dentistry have changed in human medicine with should be made to alert veterinary practitioners that antimicrobials now only recommended for dental pro- combined topical and systemic antimicrobial therapy cedures performed on patients at a high risk for cardiac should only be necessary in complicated cases where disease, to mitigate against the risk of infective endo- there is middle ear involvement with vestibular or facial carditis.24 In addition, these recommendations are now nerve dysfunction and especially when there is not in line with current accepted veterinary practice, osteomyelitis of the tympanic bulla. The use of systemic which does not recommend the use of prophylactic therapy alone is less likely to achieve the concentrations antimicrobial therapy for routine dental procedures.25 at the site of infection required to eliminate the patho- This suggests that further study of the need for antimi- gen and prevent development of resistance.22 The intro- crobial therapy after dental procedures is warranted in duction of the AIDAP therapeutic guidelines, after this veterinary medicine. survey, may have improved veterinary prescribing in Agreement with AIDAP guidelines was used as an this area and ongoing monitoring of prescribing prac- indicator of gold standard therapy in this survey. The tices is warranted. guidelines were introduced in 2013, 2 years after the The Australian Veterinary Association has also survey was conducted, so some changes in usual ther- recently recommended that antimicrobials with a high apy may have occurred after the survey was conducted importance rating such as 3rd-generation cephalospor- while guidelines were being generated. However, there ins and fluoroquinolones “should be used only when were no significant introductions of new antimicrobial other options are unavailable and wherever possible drugs into the Australian companion animal market only after susceptibility testing has been completed”.23 over this period and use of an indicator of best practice Several drugs with a high importance require autho- will allow for further investigation of factors confound- rization before administration in human medicine in ing prescribing habits. Disagreement with guidelines Australia.9 Half of the population of veterinarians was mainly due to drug selection and duration of ther- that participated in this survey accounted for 88% of apy. This was due to both overuse of therapy and lack the usage of antimicrobials of high importance rating. of recognition and treatment of severe sepsis. There was The factors that influence these prescribing habits no difference in the prescribing habits between could not be elucidated in this study. There was no recently graduated veterinarians compared to older difference in population or socioeconomic variables veterinarians. based on postcode, or in year of graduation of There are several features of this study that may have prescribers. The concurrent low use of directed antimi- influenced the results. Nonrandom, self-selection of sur- crobial therapy in low users of antimicrobials of vey respondents can result in selection bias; for exam- high importance rating may suggest that these practi- ple, veterinarians more aware or interested in tioners have a client base that is less willing to invest antimicrobial stewardship may have been more likely to

7 1106 Hardefeldt et al respond. Recall bias can occur with retrospective ques- pets and assessment of associated risk markers using a generalized tionnaire-based surveys. In order to minimize this, gen- linear mixed model. Prev Vet Med 2014;117:28–39. eric hypothetical scenarios were posed rather than 4. Davis JA, Jackson CR, Fedorka-Cray PJ, et al. Carriage of asking clinicians to recall specific cases. Prevarication methicillin-resistant Staphylococci by healthy companion animals bias was also possible. Given that this was not an in the US. Lett Appl Microbiol 2014;59:1–8. 5. Loeffler A, Pfeiffer DU, Lindsay JA, et al. Prevalence of and anonymous survey, respondents may have felt pressured risk factors for MRSA carriage in companion animals: A survey to respond in a certain way resulting in under- or over- of dogs, cats and horses. Epidemiol Infect 2011;139:1019–1028. reporting of prescribing practices and use of C & S test- 6. Boost MV, O’Donoghue MM, James A. Prevalence of Sta- ing. While bias may have affected responses to some phylococcus aureus carriage among dogs and their owners. Epi- questions, it is the authors’ opinion, given the consis- demiol Infect 2008;136:953–964. tency with other studies and our clinical experience, we 7. Grinberg A, Kingsbury DD, Gibson IR, et al. Clinically can have a reasonable level of confidence in the external overt infections with methicillin-resistant Staphylococcus aureus in validity of these findings. animals in New Zealand: A pilot study. N Z Vet J 2008;56:237– In conclusion, this survey has shown that generally 242. the choice to use antimicrobials by Australian veteri- 8. Watson ADJ, Maddison JE. Systemic antibacterial drug use in dogs in Australia. Aust Vet J 2001;79:740–746. narians is appropriate and that in the majority of 9. Australian Strategic and Technical Advisory Group on scenarios, antimicrobials with a low or medium impor- Antimicrobial Resistance. Importance rating and summary of tance rating are used. The use of antimicrobials with a antibacterials used in human health in Australia. Commonwealth high importance rating, particularly fluoroquinolones of Australia; 2015. Available at: http://www.health.gov.au/interne in dogs and 3rd-generation cephalosporins in cats, as t/main/publishing.nsf/Content/ohp-amr.htm. Accessed August 26, an empirical therapeutic choice warrants further inves- 2016. tigation now that the AIDAP guidelines have been 10. Doohoo I, Martin W, Stryhn H. Veterinary Epidemiology introduced. Research, 2nd ed. Charlottetown, CA: VER Inc; 2009. 11. Australian Bureau of Statistics. 2033.0.55.001 census of population and housing: Socio-economic indexes for areas (SEIFA), Australia; 2011. Available at: http://www.abs.gov.au/ 2011. Accessed November 11, 2016. Footnote 12. Holloway S, Trott DJ, Shipstone M, et al. Antibiotic pre- scribing detailed guidelines. Australasian Infectious Diseases Advi- a StataCorp, 2013, Stata Statistical Software: Release 13, Stata- sory Panel; 2013. Available at: http://www.ava.com.au/sites/defa Corp LP, College Station, TX ult/files/AVA_website/pdfs/AIDAP guidelines.pdf. Accessed August 12, 2016. 13. Therapeutic guidelines Limited. eTG complete; 2016. Avail- able at: http://www.tgldcdp.tg.org.au/. Accessed October 14, 2016. 14. Pleydell EJ, Souphavanh K, Hill KE, et al. Descriptive Acknowledgments epidemiological study of the use of antimicrobial drugs by com- Grant support: This project was funded by Zoetis panion animal veterinarians in New Zealand. N Z Vet J Animal Health. S.A. Holloway, D.J. Trott, M. Ship- 2012;60:115–122. 15. Mateus A, Brodbelt DC, Barber N, et al. Antimicrobial stone, V. Barrs, R. Malik, and M. Burrows, and J. usage in dogs and cats in first opinion veterinary practices in the Morton received financial remuneration for their partic- UK. J Small Anim Pract 2011;52:515–521. ipation in AIDAP. 16. Murphy CP, Reid-Smith R, Boerlin P, et al. Out-patient Conflict of Interest Declaration: S.A. Holloway, D.J. antimicrobial use in dogs and cats for new disease events from Trott, M. Shipstone, V. Barrs, R. Malik, M. Burrows, community companion animal practices in Ontario. Can Vet J and J. Morton received financial remuneration for their 2012;53:291–298. participation in AIDAP. Given that this was a cross- 17. Defauw PA, Van de Maele I, Duchateau L, et al. Risk fac- sectional survey, respondents were self-selected, and the tors and clinical presentation of cats with feline idiopathic cystitis. primary author and last author are not part of AIDAP, J Feline Med Surg 2011;13:967–975. the risk of bias is low. 18. Kruger JM, Osborne CA, Goyal SM, et al. Clinical evalua- tion of cats with lower urinary tract disease. J Am Vet Med Assoc Off-label Antimicrobial Declaration: Cefovecin and 1991;199:211–216. fluoroquinolones were used off-label. 19. Buffington CA, Chew DJ, Kendall MS, et al. Clinical evalu- ation of cats with nonobstructive urinary tract diseases. J Am Vet References Med Assoc 1997;210:46–50. 20. Burke S, Black V, Sanchez-Vizcaino F, et al. Use of cefove- 1. Jiang X, Yang H, Dettman B, et al. Analysis of fecal micro- cin in a UK population of cats attending first-opinion practices as bial flora for antibiotic resistance in ceftiofur-treated calves. Food- recorded in electronic health records. J Feline Med Surg 2016; borne Pathog Dis 2006;3:355–365. Epub ahead of print. 2. Rentala M, Lahti E, Kuhalampi J, et al. Antimicrobial 21. Wetzstein HG. Comparative mutant prevention concentra- resistance in Staphlococcus spp., Escherichia coli and Enterococcus tions of pradofloxacin and other veterinary fluoroquinolones indi- spp. in dogs given antibiotics for chronic dermatological disorders cate differing potentials in preventing selection of resistance. compared with non-treated control dogs. Acta Vet Scand Antimicrob Agents Chemother 2005;49:4166–4173. 2004;45:37–45. 22. Blondeau JM. New concepts in antimicrobial susceptibility 3. Leite-Martins LR, Mahu MI, Costa AL, et al. Prevalence of testing: The mutant prevention concentration and mutant selection antimicrobial resistance in enteric Escherichia coli from domestic window approach. Vet Dermatol 2009;20:383–396.

8 Antimicrobial Prescribing by Australian Veterinarians 1107

23. Australian Veterinary Association. Veterinary use of antibi- Supporting Information otics critical to human health; 2014. Available at: http://www. ava.com.au/. Accessed October 24, 2016. Additional Supporting Information may be found 24. Wilson W, Taubert KA, Gewitz M, et al. Prevention of online in the supporting information tab for this article: infective endocarditis: Guidelines from the American Heart Associ- ation. J Am Dent Assoc 2008;139:S3–S24. 25. British Small Animal Veterinary Association. PROTECT; Appendix S1. Questionnaire. 2016. Available at: http://www.bsava.com/Resources/PROTECT. aspx. Accessed August 22, 2016.

9 BACK TO TABLE OF CONTENTS

Chapter 4:

DETAILED ANTIMICROBIAL USE BY COMPANION ANIMAL VETERINARIANS

60

Contents lists available at ScienceDirect

Veterinary Microbiology

journal homepage: www.elsevier.com/locate/vetmic

Antimicrobials used for surgical prophylaxis by companion animal veterinarians in Australia

Laura Y. Hardefeldta,⁎, Glenn F. Browninga, Karin Thurskyb, James R. Gilkersona, Helen Billman-Jacobea, Mark A. Stevensona, Kirsten E. Baileya a Asia-Pacific Centre for Animal Health, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria, Australia b National Centre for Antimicrobial Stewardship, Peter Doherty Institute, Grattan St., Parkville, Victoria Australia

ARTICLE INFO ABSTRACT

Keywords: Antimicrobials are widely used in veterinary practices, but there has been no investigation into the classes of Antimicrobial antimicrobials used or the appropriateness of their use in surgical prophylaxis. Antimicrobial usage guidelines Antibiotic were published by the Australian Infectious Disease Advisory Panel (AIDAP) in 2013, but there has been no Stewardship investigation of compliance with them. This study aimed to investigate antimicrobial use for surgical prophylaxis Resistance in companion animal practice and assess compliance with AIDAP guidelines for selected conditions by Veterinary conducting a cross-sectional study of antimicrobial usage patterns of Australian veterinarians using an online Surgery questionnaire. Information solicited included: details of the respondent, the frequency with which antimicrobials were used for specific surgical conditions (including dose and duration) and practice antimicrobial use policies and sources of information about antimicrobial drugs and their uses. A total of 886 members of the Australian veterinary profession completed the survey. Few (22%) reported that their practice that had an antimicrobial use policy. Generally, the choice of antimicrobial drug was appropriate for the given surgical conditions. There was poor compliance with AIDAP guidelines for non-use of antimicrobials for routine neutering. Veterinarians caring solely for companion animals had higher odds of optimal compliance with guidelines than veterinarians in mixed species practices (OR 1.4, 95%CI 1.1–1.9). Recent graduates (> 2011) had lower odds of compliance than older graduates (OR 0.8, 95%CI 0.6–0.9). The findings suggest that antimicrobial use guidelines need to be expanded and promoted to improve the responsible use of antimicrobials in small animal practice in Australia.

1. Introduction use generally cannot be accurately traced to determine patterns of usage in individual species, except in rare circumstances, such as Antimicrobial use in humans and animals generates selective intramammary therapies, where use is largely limited to the treatment pressure that selects for antimicrobial resistance in bacterial popula- of clinical and subclinical mastitis in dairy cows. While antimicrobials tions (Jiang et al., 2006; Leite-Martins et al., 2014; Rentala et al., 2004). are used widely in veterinary practice, there has been limited investiga- With the growing threat of multiple resistant bacteria in medical tion of the classes of antimicrobials used and the appropriateness of hospitals, the community and in animals, there is an increasing focus dose rates and duration of therapy, and compliance with guidelines on veterinary antimicrobial usage (The Review on Antimicrobial have not been assessed. Resistance, 2016). Companion animals, and their owners, can share The Australian Strategic and Technical Advisory Group on skin and gut microbiota through direct contact, with exchange of Antimicrobial Resistance (ASTAG) issued an importance rating and antimicrobial resistant bacteria and resistance genes both possible summary of antibacterial drugs used in human health in Australia in (Ishihara et al., 2010; Platell et al., 2011; Walther et al., 2012; Weese 2015 (Australian Strategic and Technical Advisory Group on et al., 2006). Data on veterinary antimicrobial use in Australia is limited Antimicrobial Resistance, 2015). Those given a high importance rating to the periodic reports of the Australian Pesticides and Veterinary included piperacillin-tazobactam, ticarcillin-clavulanate, the 3rd and Medicines Authority, which reports gross antimicrobial use in the 4th generation cephalosporins, aztreonam, tigecycline, vancomycin, veterinary and agricultural sectors (Australian Pesticides and teicoplanin, amikacin, the streptogramins, fluoroquinolones and rifam- Veterinary Medicines Authority, 2014). However, this antimicrobial picin. These antimicrobials should be treated as third line therapies,

⁎ Corresponding author. E-mail address: [email protected] (L.Y. Hardefeldt). http://dx.doi.org/10.1016/j.vetmic.2017.03.027 Received 25 January 2017; Received in revised form 21 March 2017; Accepted 22 March 2017

61 L.Y. Hardefeldt et al. and thus should only be used where culture and susceptibility testing or graduation, the number of veterinarians employed in the practice, the other compelling clinical evidence indicates their use. In Australia, the practice location (rural or metropolitan), the practice type (general, only high importance antibacterial classes with formulations registered emergency, referral, university or government), the postcode, state, age for use in dogs and cats are the 3rd generation cephalosporins and group, position in practice (partner, associate, locum/casual), gender, fluoroquinolones. and post-graduate qualifications of the respondent. Survey sections 2, 3 In 2013 the Australian Infectious Disease Advisory Panel (AIDAP) and 4 required respondents to indicate the frequency (always, fre- introduced antimicrobial guidelines for a limited number of small quently, sometimes, rarely, never, or do not perform) with which they animal conditions (Holloway et al., 2013). The British Small Animal used antimicrobials for specific surgical conditions for each of three Veterinary Association (BSAVA) has also released more general anti- categories of practice (companion animal, bovine, equine). microbial use guidelines for surgical prophylaxis (British Small Animal Respondents were only required to complete those sections relevant Veterinary Association, 2016). Both these guidelines recommend to the species they treated. Companion animal practice scenarios against prophylactic antimicrobial therapy for routine neutering of included spaying, castration, femoral head and neck resection, femoral dogs and cats, and for routine, uncomplicated dental procedures. The fracture repair with a pin, exploratory laparotomy with enterotomy, BSAVA guidelines recommend perioperative antimicrobial treatment non-ulcerated dermal mass removal and dental prophylaxis without when surgery is prolonged (> 1.5 h) or involves implants, for debili- extractions on an otherwise healthy animal. The dose, time of anti- tated or immunosuppressed patients, where infections would be microbial initiation and duration of therapy were also requested, as catastrophic (e.g. in central nervous system surgery), where there is were other procedures associated with prescribing, such as whether an obvious break in asepsis, for all bowel surgery, for dental procedures animals were routinely weighed. Respondents were also asked what where there is periodontal disease, and for contaminated wounds or influenced their decision to use antimicrobials. The final section asked pre-existing infections (British Small Animal Veterinary Association, about practice antimicrobial use policies and sources of information 2016). While guidelines provide a necessary first step towards the about antimicrobials and their uses. All questions were closed except implementation of veterinary antimicrobial stewardship in Australia, where the specific antimicrobial was listed, where generic and trade there has been no investigation of compliance with these guidelines names were accepted. The survey was pre-tested at the 2016 Australian since their introduction. The AIDAP guidelines provided advice for two Veterinary Association (AVA) Victorian Division conference in surgical conditions; routine neutering and dental prophylaxis. There is a Melbourne on 5th March. The survey took 5–20 min to complete. No need to identify additional areas of antimicrobial use in veterinary changes were made to the survey following the initial testing. practice in Australia that require targeted education and development Data were downloaded from the survey software to spreadsheets of guidelines. Similarly, there is no evidence about the uptake or (Microsoft Office Excel, 2016). The entire small animal section of the compliance with the BSAVA guidelines in the United Kingdom or in survey had to be completed by each respondent to be included in the Australia. Antimicrobial use for surgical prophylaxis has been an area in analysis. Descriptive statistics were computed with percentages being human medicine where application of guidelines and monitoring has reported as a proportion of the total respondents answering a particular led to more appropriate antimicrobial therapy (Nelson et al., 2009). If question. Where respondents reported that they did not perform a similar initiatives are to be successful in veterinary medicine, there specific type of surgery, these results were excluded from the analysis needs to be a method for monitoring compliance and, most importantly, for that particular question. As the numbers of respondents varied by a method for providing this information back to the veterinary state or territory, the data were analysed using sampling weights to profession. provide an estimate of the inverse probability of a veterinarian’s The aim of this study was to investigate self-reported antimicrobial involvement in the survey, WH, as follows: use in a range of surgical conditions in small animal practice in N Australia and to assess compliance with AIDAP and BSAVA guidelines. WH = n 2. Methods Where N is the number of registered veterinarians in each state or territory in 2016, and n is the number of veterinarians who completed A cross-sectional study of antimicrobial usage by Australian veter- the survey from each state. Throughout this paper, all profession level inarians was conducted in 2016. The source population comprised the data are described using adjusted values based on survey design, 10,000 registered veterinarians in each of the states and territories of sampling weights and finite correction factors. Proportions of ques- Australia on 3 March 2016. The eligible population comprised those tionnaire responses are reported as unadjusted counts. registered veterinarians who were working in clinical practice at time A binary logistic regression model was used to identify individual of completion of the survey. The study population comprised those veterinarian-level characteristics that were associated with appropriate veterinarians who responded to requests to complete the questionnaire antimicrobial usage. The explanatory variables assessed in the model at the Australian Cattle Veterinarians conference (Uluru, March 2016), included the type of practice in which the respondent worked (mixed Australian Veterinary Association Conference (Adelaide, June 2016), species, companion animal, large animal), the practice location (rural, Bain-Fallon Equine Veterinary Conference (Melbourne, July 2016) and urban), their year of graduation (those who graduated up to and the Australian Small Animal Veterinary Association conference (Gold including 2011 and those who graduated after 2011), their gender, Coast, August 2016). In addition, announcements were made using their position in the practice (owner-partner, associate, casual-locum), social media, and professional organisation and regulatory agency the size of the veterinary practice (one or two full time veterinarians, correspondence (Australian Veterinary Association eline newsletter, more than two full time veterinarians), whether or not the respondent and state Veterinary Board websites and emails). had postgraduate qualifications, and the presence or absence of a Sample size calculations were performed to determine the number practice antimicrobial use policy. The outcome of interest was a of respondents required to make appropriate inferences from the proportion, where the numerator was the count of questions where survey. To be 95% certain that our estimate of the population the respondent was compliant with AIDAP or BSAVA guidelines and the prevalence of veterinarians using a given class of antimicrobial was denominator was the total number of scenarios answered in the survey. within 5% of the true population prevalence, a total of 384 completed Unconditional associations between each of the hypothesised ex- surveys were required. planatory variables and the outcome of interest were examined using The survey was created online using the survey software RedCap® odds ratios. Explanatory variables with unconditional associations for and consisted of 5 sections (available as Supplementary material). The which P < 0.20 (2-sided) were selected for multivariable modelling. initial section asked for demographic information, such as year of For the multivariable model the outcome of interest was parameterised

62 L.Y. Hardefeldt et al. as a function of the explanatory variables with unconditional associa- 15%; P = 0.014) and personal experience (mean difference, 9.3%; 95% tions regarded as significant at P < 0.20, as described above. CI, 1.6% to 17%; P = 0.01) as sources of information. Explanatory variables that were not significant were then removed Most respondents weighed dogs (99%; 95% CI, 98% to 100%) and from the model one at a time, beginning with the least significant, until cats (97%; 95% CI, 96% to 99%) prior to surgery. The five categories the estimated regression coefficients for all explanatory variables indicating the frequency of antimicrobial use for each surgical condi- retained were significant at an alpha level of less than 0.05. tion were combined into three groups (always/frequently, sometimes/ Explanatory variables that were excluded at the initial screening stage rarely and never). Routine dental procedures had the least antimicro- were tested for inclusion in the final model and were retained in the bial use (never used by 67% of respondents). Routine neutering had model if their inclusion changed any of the estimated regression moderate antimicrobial use (spay, never used by 56% of respondents; coefficients by more than 20%. Biologically plausible two-way interac- castration, never used by 65% of respondents). Removal of a non- tions were tested and none were significant at an alpha level of 0.05. ulcerated dermal mass similarly had moderate antimicrobial use This research was approved by the University of Melbourne Faculty (antimicrobials were never used by 50% of respondents). Both ortho- of Veterinary and Agricultural Sciences Human Ethics Advisory Group paedic conditions (femoral head and neck resection and repair of a under Approval No. 1646102. femoral fracture) and a clean-contaminated surgery (exploratory lapar- otomy) had high levels of antimicrobial use (Fig. 1). 3. Results Overall, the most frequently prescribed antimicrobial class in this survey was the aminopenicillins (54%), predominately in combination A total of 886 members of the Australian veterinary profession with clavulanic acid (74%), followed by cephalosporins (37%). There completed the survey. Of these, 721 completed the companion animal was a very low prevalence of high importance rating antimicrobial use section. All states and territories were represented, as were recent and (3.5%), within which use of enrofloxacin (59%) and 3rd generation older graduates. Practitioners only seeing companion animal repre- cephalosporins (37%) predominated. There were significant differences sented 66% (95% CI, 62% to 69%) of respondents who completed the in antimicrobial use between scenarios (Fig. 2). There was no sig- companion animal section, while 34% (95% CI, 31% to 38%) treated a nificant difference between antimicrobials used for spays or castrations, mixture of animal species. Use of antimicrobial use guidelines were so these categories were combined as neutering. There was no reported by 25% (178 of 721) respondents. A minority of veterinarians significant difference between the orthopaedic condition using an reported that they worked in a practice that had an antimicrobial use implant and that without, so these were combined as orthopaedic policy (166 of 721, 23%). A wide range of sources of information were conditions. Aminopenicillins were used more frequently for neutering reported as having influenced the veterinarian’s decision to use (91%), dermal mass removal (77%), dental prophylaxis (65%) and antimicrobials for surgical prophylaxis, with no single source of exploratory laparotomy (49%) than for orthopaedic conditions (37%, information predominating. Veterinarians reported using continuing P < 0.001). Cephalosporins were the class of antimicrobial most education (46%), experience (41%), an antimicrobial use policy (38%), frequently used for orthopaedic conditions (58%). High importance colleagues (37%), textbooks (37%), scientific literature (36%), online rating antimicrobials exceeded 1% only in exploratory laparotomy, resources (29%), undergraduate course notes 29%), MIMS (24%), the with fluoroquinolone use representing 5.0% of overall antimicrobial label (24%), practice guidelines (13%), and the practice policy (10%). use in this category. Surgical prophylaxis with multiple classes of Having an antimicrobial use policy changed the sources of information antimicrobials was most frequently reported for exploratory laparot- reported by clinicians, with those respondents having an antimicrobial omy (36%) and orthopaedic conditions (femoral head and neck use policy reportedly using practice guidelines more frequently than resection 15%, femoral fracture repair 19%). Usually, the additional those that did not (20% vs 12%, respectively; mean difference, 8.7%; class reflected a switch from parenteral to oral administration for all 95% CI, 2.6% to 15%; P = 0.009), the practice policy more frequently three scenarios (femoral head and neck resection 73%, femoral fracture (21% vs 7.6%; mean difference, 14%; 95% CI, 7.8% to 20%; repair 72%, exploratory laparotomy 60%), with the initial parenteral P < 0.001), and antimicrobial use guidelines more frequently (49% antimicrobial only given peri-operatively. When combined antimicro- vs 36%; mean difference, 12%; 95% CI, 4.7% to 20%; P = 0.008). bials were both parenteral, the most common combination was 1st/2nd Practitioners with an antimicrobial use policy in their workplace also generation cephalosporins and amoxycillin/clavulanate for both femor- reported using the scientific literature more frequently than those that al head and neck resection and femoral fracture (81% and 75%, did not have a policy in place (41% vs 35%; mean difference, 10%; 95% respectively). For exploratory laparotomy, the most common combina- CI, 2.4% to 18%; P = 0.004). Clinicians that did not have an anti- tion was metronidazole with 1st or 2nd generation cephalosporins or microbial use policy in their workplace were more likely to report using amoxycillin/clavulanate (53%), followed by 1st or 2nd generation undergraduate course notes (mean difference, 8.2%; 95% CI, 1.3% to cephalosporins with amoxycillin/clavulanate (20%), and amoxycillin/

Fig. 1. Frequency of reported antimicrobial usage for seven clinical scenarios.

63 L.Y. Hardefeldt et al.

Fig. 2. Proportions of antimicrobial usage, by class, for surgical scenarios.* HIRA: High importance rating antimicrobial. ** LIRA: Low importance rating antimicrobial. clavulanate with fluoroquinolones (19%). Topical and intraperitoneal plasma antimicrobial concentrations at the time of surgery and 32% to antimicrobial therapy was rarely reported (topical – femoral head and 38% prescribing antimicrobial therapy for longer than the guideline neck resection seven instances, femoral fracture repair six instances, recommendations (Fig. 4). Suboptimal compliance in these scenarios exploratory laparotomy two instances; intraperitoneal – exploratory was common, with the majority of respondents selecting the appro- laparotomy two instances). priate antimicrobial agent (exploratory laparotomy 57%, femoral Compliance with AIDAP and BSAVA guidelines was evaluated for fracture repair 66% and femoral head and neck resection 63%). the use or non-use of antimicrobials, drug choice, duration of therapy An individual’s overall optimal compliance was calculated as the and overall agreement. Compliance was classed as optimal if the proportion of optimally compliant scenarios out of the total number of therapy complied with all recommendations, suboptimal if the correct scenarios completed by that individual in the survey. There was marked drug choice was made but duration or timing and route of administra- variation between individual respondents, ranging from 0% to 100% tion were not compliant, and non-compliant if drug choice was optimal overall compliance. After adjusting for the effect of practice inappropriate, or if an antimicrobial was administered when none location (rural versus metropolitan), time since graduation, gender, role was recommended by the guidelines. The proportion of veterinarians in the practice, the acquisition of post-graduate qualifications and the reporting optimal compliance with guidelines for the different surgical presence or absence of a practice antimicrobial use policy, the odds of scenarios ranged from 12% to 67%. There was a moderate level of compliance was 1.4 (95% CI, 1.1–1.9) times greater for companion optimal compliance for routine dental procedures, castration, spaying animal practitioners than for mixed species practitioners. The odds of and dermal mass removal (67%, 65%, 56% and 50%, respectively), compliance for recent graduates (respondents who graduated after whereas the level of optimal compliance for orthopaedic conditions and 2011) was 0.8 (95% CI, 0.6–0.9) times that of graduates who graduated exploratory laparotomy was low (exploratory laparotomy 12%, femoral in 2011 or earlier. The odds of compliance for respondents working in fracture repair 20%, and femoral head and neck resection 28%) (Fig. 3). practices with more than two full time veterinarians was 1.4 (95% CI, The level of optimal compliance was low in these scenarios because of 1.1–1.9) times that of respondents working in one- or two-veterinarian inappropriate timing or route of therapy, or inappropriate duration of practices (Table 1). The lower optimal compliance in recent graduates therapy. For all three conditions, timing and route of administration was due to inappropriate drug choice and timing of antimicrobial was the least appropriate, with fewer than 25% of respondents administration in FF, FHNR and EXLAP. The odds of compliance were reporting a timing or route of therapy that would allow for effective lower for respondents from New South Wales (NSW), South Australia

Fig. 3. Proportions of veterinarians reporting optimal or suboptimal compliance with AIDAP and BSAVA guidelines for prophylactic therapy for different surgical scenarios.

64 L.Y. Hardefeldt et al.

Fig. 4. Proportions of veterinarians reporting compliance with guidelines for choice of antimicrobial drug, the timing and route of administration or the duration of therapy for different surgical scenarios.

Table 1 Estimated regression coefficients and their standard errors from a logistic regression model of risk factors for antimicrobial usage compliance.

Variable Compa Scenb Coefficient (SE) t P OR (95%CI)

Interceptc 1733 4401 −0.032 (0.346) −0.09 0.926 0.96 (0.49, 1.91)

Type of practice Mixed species 492 1525 Reference 1.00 Mainly companion animal 1241 2876 0.339 (0.142) 2.38 0.018 1.40 (1.06, 1.86)

Practice location Rural 615 1815 Reference 1.00 Metropolitan 1118 2586 −0.127 (0.137) −0.93 0.353 0.88 (0.67, 1.15)

Graduation ≤ 2011 452 1341 Reference 1.00 > 2011 1204 2868 −0.282 (0.112) −2.52 0.012 0.75 (0.61, 0.94)

Gender Male 558 1385 Reference 1.00 Female 1175 3016 −0.074 (0.100) −0.74 0.457 0.93 (0.76, 1.13)

Position in practice Owner/partner 1175 3016 Reference 1.00 Associate 1074 2708 0.156 (0.126) 1.24 0.214 1.17 (0.91, 1.50) Casual/locum 131 322 −0.005 (0.166) −0.03 0.975 0.99 (0.72, 1.38) Other 62 188 −0.152 (0.259) −0.59 0.557 0.86 (0.52, 1.43)

Size of practice 1 or 2 vets 859 2404 Reference 1.00 > 2 vets 1481 3608 0.355 (0.135) 2.64 0.009 1.43 (1.10, 1.86)

Postgraduate qualifications No 1305 3405 Reference 1.00 Yes 428 996 0.153 (0.107) 1.42 0.155 1.17 (0.94, 1.44)

Antimicrobial use policy No 1368 3443 Reference 1.00 Yes 365 958 −0.136 (0.111) −1.23 0.220 0.87 (0.70, 1.09)

State of practice ACT 36 62 Reference 1.00 NSW 371 936 −0.627 (0.204) −3.08 0.002 0.53 (0.36, 0.80) NT 22 47 −0.242 (0.441) −0.55 0.583 0.79 (0.33, 1.87) SA 120 293 −0.666 (0.261) −2.56 0.011 0.51 (0.31, 0.86) QLD 273 813 −0.919 (0.206) −4.45 < 0.001 0.40 (0.27, 0.60) TAS 39 105 −0.623 (0.405) −1.54 0.125 0.54 (0.24, 1.19) VIC 504 1303 −0.678 (0.199) −3.41 0.001 0.51(0.34, 0.75) WA 368 842 −0.598 (0.210) −2.85 0.005 0.55 (0.36, 0.83)

Practice type General Practice 1547 3952 Reference 1.00 Emergency 71 164 −0.031 (0.241) −0.13 0.898 0.97 (0.60, 1.56) Referral 74 191 −0.328 (0.236) −1.39 0.165 0.72 (0.45, 1.14) University 27 73 −0.521 (0.378) −1.38 0.168 0.59 (0.28, 1.25)

SE: Standard error; OR: Odds Ratio; CI: Confidence interval. a Number of optimally compliant scenarios. b Total number of scenarios answered. c Baseline compliance is adjusted for sampling fraction.

5 L.Y. Hardefeldt et al.

(SA), Queensland (QLD), Victoria (VIC) and Western Australia (WA) quent use of the combination of amoxycillin/clavulanate and metroni- than for respondents from the Australian Capital Territory (ACT) (range dazole in exploratory laparotomy was also unexpected, as resistance to 45–60%). amoxycillin/clavulanate in anaerobes appears to be rare in veterinary practice (Gibson et al., 2008; Lawhon et al., 2013). 4. Discussion The wide range of sources used for information on antimicrobial prescribing for surgical prophylaxis likely reflects that there is no This is the first survey to investigate antimicrobial use for surgical central, high quality, source of information available to veterinarians. prophylaxis in the veterinary profession. In addition, it is the first to The frequent reporting of experience as an information source by over report on compliance with antimicrobial use guidelines in veterinary 40% of respondents supports this interpretation. Encouragingly, con- medicine. Antimicrobial usage for surgical prophylaxis was relatively tinuing education was also a common source of information (46% of commonplace, but compliance with the published guidelines was respondents). Those veterinarians working in practices with an anti- variable. Optimal compliance was moderate for commonly performed microbial use policy reported using practice guidelines, an antimicro- routine surgeries, such as de-sexing and dental surgery, but was low for bial use policy and practice policies more frequently than those more specialised surgeries, such as orthopaedic surgeries and explora- veterinarians working in practices without a policy, but there was no tory laparotomies. difference in compliance between these groups of veterinarians. This Two studies have been conducted into the antimicrobial prescribing may suggest that practice antimicrobial use policies are not compliant practices of Australian veterinarians (14, author’s unpublished results). with the AIDAP and BSAVA guidelines. In addition, while the AIDAP The AIDAP group surveyed the veterinary profession on the use of guidelines were published in 2013, only 25% of respondents reported antimicrobial agents for a range of medical and dermatological using guidelines as a source of information. The low uptake of the scenarios, but also included routine desexing and dental prophylaxis guidelines needs to be investigated. A widely available, reputable and (Hardefeldt et al. unpublished). Antimicrobials were reportedly used well publicised source of information may aid in improving antimicro- for routine desexing by only 25% of respondents in the AIDAP study, in bial prescribing in companion animal practice. contrast to the findings of the survey reported here, in which 44% of Optimal compliance with AIDAP and BSAVA guidelines varied respondents reported some use of antimicrobials for spaying and 35% across scenarios. However suboptimal compliance was high in those of respondents reported some use of antimicrobials for castrations. The scenarios for which antimicrobial prophylaxis was recommended. The AIDAP surveyed was predominately of companion animal veterinar- major reasons for non-compliance were use of a time or route of ians, whereas this study had a much broader mix of companion animal administration that would be expected to have precluded development and mixed species veterinarians, which may explain the differences. of appropriate plasma levels of the antimicrobial at the time of surgery, Interestingly, while antimicrobials were reported to be used more or therapy for longer than the maximum of 24 h recommended in the frequently by respondents to this survey than in the AIDAP survey, guidelines. antimicrobials were reported to be used less frequently for routine The results of this study show that companion animal practitioners dental prophylaxis by respondents in this survey (33% vs 90%). This were more compliant with guidelines than mixed species veterinarians. may reflect differences in the survey questions, as the AIDAP survey In addition, practitioners employing 2 or more veterinarians were more included dental extractions, or a change in prescribing practices, with compliant. This combination of findings may reflect the increased increased recognition that the risk of infectious endocarditis following likelihood of a single species focus of veterinarians in larger practices dental procedures is likely to be very low, as is the case in human and greater access to continuing education for that single species. dentistry (Wilson et al., 2008). Veterinarians from larger practices may also share knowledge within Overall, more than 90% of the antimicrobials reported to be used for the practice to optimise treatment choices. Continuing education was surgical prophylaxis in this survey have low or medium importance cited as a source of information on antimicrobial prescribing by almost ratings. These findings are consistent with previous studies from half of the respondents in this survey. Recent graduates were less Australia (AIDAP), Canada (Murphy et al., 2012), the United Kingdom compliant than older graduates. It is possible this is due to poor (Mateus et al., 2011) and New Zealand (Pleydell et al., 2012), in which university teaching, but it is more likely that peer pressure from aminopenicillins and 1st generation cephalosporins were reported to be colleagues, practice protocols and the lack of confidence of new the most frequently prescribed classes of antimicrobial drugs in graduates results in them following the practices of other veterinarians veterinary medical scenarios. Antimicrobial choice varied between in their workplace. In the only other study that examined agreement surgical scenarios in this study, with aminopencillins being most with guidelines, there was no difference between companion animal frequently selected for soft tissue surgeries, while cephalosporins were only and mixed species veterinarians, or between recent and older more frequently selected for orthopaedic conditions. The BSAVA graduates (Hardefeldt et. al., unpublished). This study was of a mix of guidelines recommend either amoxycillin/clavulanate or 1st generation predominately medical and dermatological scenarios, with only 2 cephalosporins for any surgical prophylaxis, unless there is significant surgical scenarios. It may be that prescribing behavior differs between bowel leakage or likely to be involvement of anaerobic bacteria. The surgical and medical scenarios for less experienced veterinarians. The drivers for the choice of aminopenicillins or cephalosporins for surgical odds of optimal compliance was lower in all states, except the NT and prophylaxis by Australian veterinarians revealed in this survey are not TAS, where low numbers of respondents may have affected the results, clear. While the prevalence of fluoroquinolone use was very low than in the ACT. This may be early evidence of an impact from the overall, its use by 5% of respondents for an exploratory laparotomy recent initiation of an antimicrobial stewardship programme in a large with an enterotomy warrants monitoring as this use is contrary to the proportion of the companion animal practices in the ACT (Taylor, guidelines and appears unnecessary. amoxycillin/clavulanate will 2016). provide appropriate coverage unless there is gross contamination of There are several features of this study that may have influenced the the abdomen with gastrointestinal contents, in which case addition of results. Recall bias may occur with retrospective questionnaire based an aminoglycoside is recommended in the BSAVA guidelines (British surveys. Hypothetical scenarios were posed rather than asking clin- Small Animal Veterinary Association, 2016). The use of combined icians to recall specific cases in order to minimise this. Respondents therapy, in femoral head and neck resection, femoral fracture repair were self-selected in this study and many were recruited at conferences, and exploratory laparotomy, with both amoxycillin/clavulanate and 1st so selection bias may also be present. This may have biased the results generation cephalosporins was unexpected. Practitioners were probably towards practitioners who are more likely to engage in continuing reporting that they either use amoxycillin/clavulanate or 1st generation education, and have more awareness of recommended prescribing cephalosporins interchangeably, rather than both together. The fre- practices. While results were adjusted for lack of equivalence in state 6 L.Y. Hardefeldt et al. sampling, other adjustments to account for population variations were British Small Animal Veterinary Association, 2016. Protect. http://www.bsava.com/ Resources/PROTECT.aspx. not possible, as population statistics for the Australian veterinary Gibson, J.S., Morton, J.M., Cobbold, R.N., Sidjabat, H.E., Filippich, L.J., Trott, D.J., 2008. profession are largely unknown. However, all age-groups were repre- Multi-drug resistant E. coli and Enterobacter extraintestinal infection in 37 dogs. J. sented, as were rural and metropolitan veterinarians. The survey was Vet. Intern. Med. 22, 844–850. Holloway, S., Trott, D.J., Shipstone, M., Barrs, V., Malik, R., Burrows, M., 2013. Antibiotic also anonymous, to minimise response bias. Prescribing Detailed Guidelines. Australasian Infectious Diseases Advisory Panel. http://www.ava.com.au/sites/default/files/AVA_website/pdfs/AIDAP guidelines. 5. Conclusion pdf. Ishihara, K., Shimokubo, N., Sakagami, A., Ueno, H., Muramatsu, Y., Kadosawa, T., Yanagisawa, C., Hanaki, H., Nakajima, C., Suzuki, Y., Tamura, Y., 2010. Occurrence This survey has shown that, while antimicrobials are commonly and molecular characteristics of methicillin-resistant Staphylococcus aureus and used for surgical prophylaxis in Australia, the levels of poor compliance methicillin-resistant Staphylococcus pseudintermedius in an academic veterinary with AIDAP and BSAVA guidelines are relatively low. There was very hospital. Appl. Environ. Microbiol. 76, 5165–5174. Jiang, X., Yang, H., Dettman, B., Doyle, M.P., 2006. Analysis of fecal microbial flora for low use of antimicrobials with a high-importance rating for any of the antibiotic resistance in ceftiofur-treated calves. Foodborne Pathog. Dis. 3, 355–365. scenarios presented to respondents. Education is warranted to improve Lawhon, S.D., Taylor, A., Fajt, V.R., 2013. Frequency of resistance in obligate anaerobic timing of antimicrobial administration and shorten the duration of bacteria isolated from dogs, cats, and horses to antimicrobial agents. J. Clin. Microbiol. 51, 3804–3810. surgical prophylaxis. In addition, the non-use of antimicrobials for Leite-Martins, L.R., Mahu, M.I., Costa, A.L., Mendes, A., Lopes, E., Mendonca, D.M., Niza- routine clean surgery, such as desexing, should be promoted more Ribeiro, J.J., de Matos, A.J., da Costa, P.M., 2014. Prevalence of antimicrobial strongly, as should the adoption of antimicrobial use policies by resistance in enteric Escherichia coli from domestic pets and assessment of associated risk markers using a generalized linear mixed model. Prev. Vet. Med. 117, 28–39. veterinary practices in Australia. Mateus, A., Brodbelt, D.C., Barber, N., Stark, K.D., 2011. Antimicrobial usage in dogs and cats in first opinion veterinary practices in the UK. J. Small Anim. Pract. 52, 515–521. Conflicts of interest Murphy, C.P., Reid-Smith, R., Boerlin, P., Weese, J.S., Prescott, J.F., Janecko, N., McEwen, S.A., 2012. Out-patient antimicrobial use in dogs and cats for new disease events from community companion animal practices in Ontario. Can. Vet. J. 53, None. 291–298. Nelson, R.L., Glenny, A.M., Song, F., 2009. Antimicrobial prophylaxis for colorectal Funding surgery. Cochrane Database Syst. Rev. CD001181. Platell, J.L., Cobbold, R.N., Johnson, J.R., Heisig, A., Heisig, P., Clabots, C., Kuskowski, M.A., Trott, D.J., 2011. Commonality among fluoroquinolone-resistant sequence type This work was funded by the National Health and Medical Research ST131 extraintestinal Escherichia coli isolates from humans and companion animals in Council (National Centre for Antimicrobial Stewardship – Grant No. Australia. Antimicrob. Agents Chemother. 55, 3782–3787. Pleydell, E.J., Souphavanh, K., Hill, K.E., French, N.P., Prattley, D.J., 2012. Descriptive 1079625) epidemiological study of the use of antimicrobial drugs by companion animal veterinarians in New Zealand. N. Z. Vet. J. 60, 115–122. Acknowledgments Rentala, M., Lahti, E., Kuhalampi, J., Pesonen, S., Jarvinen, A.K., Saijonmaa-Koulumies, L., Honkanen-Buzalski, T., 2004. Antimicrobial resistance in Staphlococcus spp., Escherichia coli and Enterococcus spp. in dogs given antibiotics for chronic This research was funded by the National Health and Medical dermatological disorders compared with non-treated control dogs. Acta Vet. Scand. Research Council through the Centres of Research Excellence pro- 45, 37–45. Taylor, A., 2016. Antimicrobial stewardship in companion animal practice: a pilot gramme, grant no. 1079625. LYH was a recipient of an Australian programme in Canberra Australia. In: 4th Responsible Use of Antibiotics in Animals. Postgraduate Award scholarship. The Hague. . The Review on Antimicrobial Resistance, 2016. Tackling Drug-Resistant Infections: An Overview of Our Work. In: O'Neill, J. (Ed.), . http://www.amr-review.org/. Appendix A. Supplementary data Walther, B., Hermes, J., Cuny, C., Wieler, L.H., Vincze, S., Abou Elnaga, Y., Stamm, I., Kopp, P.A., Kohn, B., Witte, W., Jansen, A., Conraths, F.J., Semmler, T., Eckmanns, Supplementary data associated with this article can be found, in the T., Lubke-Becker, A., 2012. Sharing more than friendship-nasal colonization with online version, at http://dx.doi.org/10.1016/j.vetmic.2017.03.027. coagulase-positive staphylococci (CPS) and co-habitation aspects of dogs and their owners. PLoS One 7, e35197. Weese, J.S., Dick, H., Willey, B.M., McGeer, A., Kreiswirth, B.N., Innis, B., Low, D.E., References 2006. Suspected transmission of methicillin-resistant Staphylococcus aureus between domestic pets and humans in veterinary clinics and in the household. Vet. Microbiol. 115, 148–155. Australian Pesticides and Veterinary Medicines Authority, 2014. Quantity of Wilson, W., Taubert, K.A., Gewitz, M., Lockhart, P.B., Baddour, L.M., Levison, M., Bolger, Antimicrobial Products Sold for Veterinary Use in Australia: July 2005 to June 2010. A., Cabell, C.H., Takahashi, M., Baltimore, R.S., Newburger, J.W., Strom, B.L., Tani, http://www.apvma.gov.au/. L.Y., Gerber, M., Bonow, R.O., Pallasch, T., Shulman, S.T., Rowley, A.H., Burns, J.C., Australian Strategic, Technical Advisory Group on Antimicrobial Resistance, 2015. Ferrieri, P., Gardner, T., Goff, D., Durack, D.T., 2008. Prevention of infective Importance Rating and Summary of Antibacterials Used in Human Health in endocarditis: guidelines from the American Heart Association. J. Am. Dental Assoc. Australia. Commonweath of Australia. http://www.health.gov.au/internet/main/ 139, S3–S24. publishing.nsf/Content/ohp-amr.htm.

7 BACK TO TABLE OF CONTENTS

Chapter 5:

DETAILED ANTIMICROBIAL USE BY EQUINE VETERINARIANS

6

8 Equine Veterinary Journal ISSN 0425-1644 DOI: 10.1111/evj.12709

Antimicrobials used for surgical prophylaxis by equine veterinary practitioners in Australia L. Y. HARDEFELDT†‡* , G. F. BROWNING†‡, K. THURSKY‡, J. R. GILKERSON†, H. BILLMAN-JACOBE†‡, M. A. STEVENSON† and K. E. BAILEY†‡

†Asia-Pacific Centre for Animal Health, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria, Australia ‡National Centre for Antimicrobial Stewardship, Peter Doherty Institute, Carlton, Victoria, Australia.

*Correspondence email: [email protected]; Received: 21.02.17; Accepted: 02.06.17

Summary

Background: Antimicrobials are widely used in Australian veterinary practices, but no investigation into the classes of antimicrobials used, or the appropriateness of use in horses, has been conducted. Objectives: The aim of the study was to describe antimicrobial use for surgical prophylaxis in equine practice in Australia. Study design: Cross-sectional questionnaire survey. Methods: An online questionnaire was used to document antimicrobial usage patterns. Information solicited in the questionnaire included demographic details of the respondents, the frequency with which antimicrobials were used for specific surgical conditions (including the dose, timing and duration of therapy) and practice antimicrobial use policies and sources of information about antimicrobials and their uses. Results: A total of 337 members of the Australian veterinary profession completed the survey. Generally, the choice of antimicrobial was appropriate for the specified equine surgical condition, but the dose and duration of therapy varied greatly. While there was poor optimal compliance with British Equine Veterinary Association guidelines in all scenarios (range 1–15%), except removal of a nonulcerated dermal mass (42%), suboptimal compliance (compliant antimicrobial drug selection but inappropriate timing, dose or duration of therapy) was moderate for all scenarios (range 48–68%), except for an uninfected contaminated wound over the thorax, where both optimal and suboptimal compliance was very poor (1%). Veterinarians practicing at a university hospital had higher odds of compliance than general practice veterinarians (Odds ratio 3.2, 95% CI, 1.1–8.9, P = 0.03). Main limitations: Many survey responses were collected at conferences which may introduce selection bias, as veterinarians attending conferences may be more likely to have been exposed to contemporary antimicrobial prescribing recommendations. Conclusions: Antimicrobial use guidelines need to be developed and promoted to improve the responsible use of antimicrobials in equine practice in Australia. An emphasis should be placed on antimicrobial therapy for wounds and appropriate dosing for procaine penicillin.

Keywords: horse; antimicrobial; stewardship; resistance; surgery

Introduction In Australia, there are no guidelines for antimicrobial use in equines. The British Equine Veterinary Association (BEVA) has released antimicrobial use Antimicrobial use in man and animals generates selective pressure that guidelines including guidance for antimicrobial use in surgical prophylaxis selects for increased levels of antimicrobial resistance in bacterial [16]. These guidelines recommend administration of penicillin pre- and populations [1–3]. With the growing threat of antimicrobial resistant post-operatively for 24 h for clean surgeries, penicillin and gentamicin pre- bacteria in medical hospitals, the community and in animals, there is an and post-operatively for 5 days for contaminated surgeries and penicillin increasing focus on veterinary antimicrobial usage [4]. Companion animals, and gentamicin pre- and post-operatively for 10 days for high risk including horses and their owners, share skin and gut microbiota through surgeries. For uncomplicated contaminated wounds, antimicrobial therapy direct contact, with exchange of antimicrobial resistant bacteria possible is not recommended. Whilst these guidelines provide a necessary first step [5–10]. Data on the amounts of antimicrobial use in animals in Australia is towards the implementation of veterinary antimicrobial stewardship, audit limited to periodic reports provided by the Australian Pesticides and and feedback are necessary to improve prescribing practices. Antimicrobial Veterinary Medicines Authority, which records the amounts of use for surgical prophylaxis has been an area in human medicine where antimicrobial drugs imported for use in the veterinary and agricultural application of guidelines and monitoring has led to more appropriate sectors [11]. While these data provide a useful starting point for improved antimicrobial therapy [17]. antimicrobial stewardship, the quantities reported cannot be stratified by The aim of this study was to investigate self-reported antimicrobial use species or production type, except for specific formulations, such as in a range of surgical conditions in equine practice in Australia and to intramammary therapies, where use is largely limited to the treatment of assess compliance with BEVA’s PROTECT-ME guidelines [16]. clinical and subclinical mastitis in dairy cows. Recent studies from Canada [12,13] and the UK [14] have concluded that inappropriate use of antimicrobials is common in horses. However, the classes of antimicrobials, Materials and methods appropriateness of drug doses and duration of therapy used by equine practitioners in surgical prophylaxis have not been comprehensively Details of the source, eligible and study group for this study are described examined. elsewhere [18]. Briefly, the eligible population comprised those registered The Australian Strategic and Technical Advisory Group on Antimicrobial veterinarians who were working in equine practice in Australia at the time Resistance (ASTAG) issued importance ratings of antibacterials used in of completion of the questionnaire (estimated to be 2960 veterinarians [19] human health in Australia in 2015 [15]. Of the antimicrobials licensed for in 2015). use in horses, third generation cephalosporins are listed as high importance by ASTAG and therefore it is recommended that they should Sample size calculations only be used where culture and susceptibility testing, or other compelling Sample size calculations were carried out to determine the number of clinical evidence, provide justification for their use. respondents required to make appropriate inferences from the survey. To

Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd

9 Antimicrobials for surgical prophylaxis L. Y. Hardefeldt et al. be 95% certain that our estimate of the population prevalence of Results veterinarians using a given class of antimicrobials was within 6% of a true population prevalence of 50%, a total of 245 completed surveys were A total of 337 members of the Australianveterinaryprofessioncompleted required. Sample size calculations were carried out assuming a 50% the survey. All states and territories were represented, as were recent and population prevalence because this provided the largest sample size older graduates. Respondents were predominately from first opinion estimate for a constant margin of error. practice (87%), with the remainder from referral and university practice Study data were collected and managed using REDCap electronic data (13%). Equine only practitioners represented 31% (95% CI, 21–40%) of capture tools [20]. The survey details are described elsewhere [18]. The respondents, while the remaining 69% (95% CI, 63–75%) treated a mixture of equine surgical scenarios included in the survey were castration, a clean species (3.0% equine and bovine, 8.5% equine and companion animals, 56% wound that is sutured, removal of a nonulcerated dermal mass, a equine, bovine and companion animals). The veterinarians completing the contaminated wound over the thorax that was not sutured and not survey served a variety of client sectors (39% pleasure horses only, 7.5% infected, an uncomplicated umbilical hernia repair, a transphyseal bridge racetrack only, 3.1% reproductive practice only and 50% crossing a mix with a lag screw and an uncomplicated eye ablation (see Supplementary of sectors). Few practices had an antimicrobial use policy (20%, 95% CI, Item 1). 11–31%). A very wide range of sources of information on antimicrobials for Data analysis surgical prophylaxis were reported, with no single source of information predominating. Practitioners reported using experience (12%), continuing a Data were downloaded from the survey software into spreadsheets . education (11%) and textbooks (11%) mostfrequently,with7%reportingthe The entire equine section of the survey had to be completed by each label as an important source of information. Having an antimicrobial use respondent to be included in the analysis of equine practitioner policy did not change the sources of information reported by clinicians, antimicrobial usage. Descriptive statistics were computed with with only 7.1% of practitioners with a policy in place reporting this as a percentages for each response, calculated as the proportion of the source of information for antimicrobial therapy. The amount of total number of respondents answering a particular question. Where contamination (33%) and surgical conditions (23%) were the most frequently respondents reported that they did not perform a specific surgery, reported influences on the decision to prescribe antimicrobials in the they were excluded from the analysis for that question. The data were surgical scenarios. analysed using sampling weights, WH, to provide an estimate of the inverse probability of a veterinarian’s involvement in the survey, as the Antimicrobial use numbers of respondents varied by state or territory, which were quantified as follows: The five categories indicating the frequency of antimicrobial use for each surgical condition were combined into three groups (always/frequently, N sometimes/rarely and never). Removal of a nonulcerated dermal mass had WH 1 ¼ n ð Þ the least antimicrobial use; antimicrobials were used always or frequently by 26% of respondents, sometimes or rarely by 34% of respondents and Where N is the number of registered veterinarians in each state or territory never by 40% of respondents. The frequency of antimicrobial use was high in 2016 and n is the number of veterinarians who completed the survey for all other types of surgery (Fig 1). from each state. Throughout this paper, all profession level data are Overall, the most frequently prescribed antimicrobial class in this survey described using adjusted values based on survey design, sampling weights was the penicillins (70%), predominantly procaine penicillin (95%). The only and finite correction factors. Proportions of questionnaire responses are other frequently prescribed antimicrobials were trimethoprim- reported as unadjusted counts. sulphonamide (16%) and gentamicin (12%). All other classes represented A logistic regression model was used to identify individual veterinarian less than 1.3% of all reported antimicrobials in the survey (Fig 2). Penicillin level characteristics associated with appropriate antimicrobial usage. The was the most commonly used antimicrobial for all the surgical scenarios explanatory variables assessed in the model included the type of (48–86% of respondents across scenarios). Use of trimethoprim- practice in which the respondent worked (mixed species, large animal, sulphonamide, or a combination therapy of penicillin and gentamicin were equine only), practice location (rural, metropolitan), year of graduation also frequently reported for all scenarios except for castration (Fig 3). (those who graduated up to and including 2011 and those who There was a very low incidence of use of antimicrobials with a high graduated after 2011), gender, position in the practice (owner-partner, importance rating (0.7%), with third generation cephalosporins being the associate, casual-locum), size of the veterinary practice (one or two full only drugs in this group with use not exceeding 1.9% of antimicrobials used time veterinarians, more than two full time veterinarians), whether or not in any one scenario. There was wide variation in duration of therapy the respondent had post-graduate qualifications and the presence or between scenarios. Castration was the only scenario in which antimicrobial absence of a practice antimicrobial use policy. The outcome of interest therapy was stopped within 24 h by most respondents (66%). Antimicrobial was a proportion, where the numerator was the count of questions in therapy was the longest for both wound scenarios, with 74 and 53% of which the respondent was compliant with BEVA guidelines and the respondents indicating they treated for 3–7daysforcontaminatedwounds denominator was the total number of scenarios answered in the survey. (not infected and left open to heal by secondary intention) and clean Generalised linear regression models were fitted and Z tests performed, wounds that were sutured, respectively. using functions within Stata v13. Unconditional associations between each of the hypothesised Optimal compliance explanatory variables listed above and the outcome of interest were Compliance with BEVA guidelines on the use or nonuse of antimicrobials, computed using the odds ratio. Explanatory variables with unconditional drug choice, dose and duration of therapy was evaluated, as well as overall associations significant at the P<0.20 level (two-sided) were selected for agreement with these guidelines. A 10% margin of error was allowed when multivariable modelling. All explanatory variables meeting this criterion assessing compliance with guidelines for dosage. Compliance was classed were entered into the multivariable model. Explanatory variables that were as optimal if the therapy complied with all recommendations, suboptimal if not significant were then removed from the multivariable model one at a the correct drug choice was made but dose, duration of therapy or timing time, beginning with the least significant, until the estimated regression of antimicrobial therapy prior to surgery were not compliant and coefficients for all explanatory variables retained were significant at an noncompliant if drug choice was inappropriate or if an antimicrobial was alpha level of less than 0.05. Explanatory variables that were excluded at administered when none was recommended by the guidelines. The the initial screening stage were tested for inclusion in the final model and frequency of optimal compliance for the different surgical scenarios ranged were retained in the model if their inclusion changed any of the estimated from 1.1% to 42%. Except for removal of a nonulcerated dermal mass, regression coefficients by more than 20%. Biologically plausible two-way optimal compliance was low for all scenarios (1.1–25%) (Fig 4). Suboptimal interactions were tested and none were significant at an alpha level of compliance was common for all scenarios, with most respondents 0.05. selecting the appropriate antimicrobial agent (36–68%) (Fig 4). Suboptimal

2 Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd

70 L. Y. Hardefeldt et al. Antimicrobials for surgical prophylaxis

TABLE 1: Estimated regression coefficients and their standard errors from a logistic regression model of risk factors for compliance with guidelines for prophylactic antimicrobial usage in surgery

Variable Comp1 Scenario2 Coefficient (SE) t P OR (95% CI)

Intercept3 281 1419 1.25 (0.539) 2.33 0.02 0.47 (0.25, 0.86) À À Type of practice General practice 207 1183 Reference Referral 35 148 0.078 (0.393) 0.20 0.9 1.08 (0.50, 2.34) University 35 81 1.160 (0.522) 2.22 0.03 3.19 (1.14, 8.91) Practice location Rural 190 1048 Reference Metropolitan 91 371 0.138 (0.268) 0.52 0.6 0.87 (0.51, 1.47) À À Graduation ≤2011 212 1017 Reference >2011 69 402 0.513 (0.267) 1.92 0.06 0.60 (0.35, 1.01) À À Gender Male 109 596 Reference Female 172 823 0.200 (0.217) 0.92 0.4 1.22 (0.80, 1.87) Position in practice Owner/partner 92 514 Reference Associate 139 736 0.098 (0.274) 0.36 0.7 1.10 (0.64, 1.89) Casual/locum 8 62 1.04 (0.621) 1.68 0.09 0.35 (0.10, 1.20) À À Other 42 107 0.371 (0.433) 0.86 0.4 1.45 (0.62, 3.40) À Size of practice ≤2 veterinarians 50 300 Reference >2 veterinarians 231 1119 0.175 (0.250) 0.70 0.5 1.19 (0.73, 1.95) Post-graduate qualifications No 184 1041 Reference Yes 97 378 0.253 (0.242) 1.05 0.3 1.28 (0.80, 2.07) Antimicrobial use policy No 236 1126 Reference Yes 45 293 0.411 (0.263) 1.56 0.1 0.66 (0.40, 1.11) À À Species treated Equine only 130 560 Reference Equine/bovine 7 51 0.170 (0.656) 0.26 0.8 0.84 (0.23, 3.06) À À Mixed 119 701 0.174 (0.287) 0.61 0.6 0.84 (0.48, 1.48) À À Small/equine 25 107 0.613 (0.420) 1.46 0.2 1.85 (0.81, 4.22) State of practice3 NSW 84 383 Reference NT 11 20 0.736 (0.429) 1.72 0.09 2.10 (0.90, 4.86) SA 12 97 0.920 (0.543) 1.70 0.09 0.40 (0.14, 1.16) À À QLD 56 253 0.117 (0.288) 0.41 0.7 0.89 (0.50, 1.57) À À TAS 9 42 0.003 (0.454) 0.01 >0.9 1.00 (0.41, 2.45) VIC 87 467 0.319 (0.248) 1.29 0.2 0.73 (0.45, 1.18) À À WA 22 157 0.673 (0.350) 1.92 0.06 0.51 (0.26, 1.01) À À SE, standard error; OR, odds ratio; CI, confidence interval. 1Number of optimally compliant scenarios. 2Total number of scenarios answered. 3Baseline compliance adjusted for sampling fraction. compliance was not evaluated for the contaminated wound scenario as inappropriate, with fewer than 30% of respondents reporting using antimicrobials were not indicated. appropriate doses of antimicrobials in all scenarios, except transphyseal bridging. Subtherapeutic dosing of penicillin accounted for the majority of Suboptimal compliance underdosing, with 68% of all procaine penicillin dose rates being lower than Optimal compliance was low due to inappropriate dose, inappropriate that recommended in the literature, whereas only 12% of gentamicin dose timing of administration of drug and/or inappropriate duration of therapy. rates were lower than those recommended in the literature. Duration of Timing of administration was the aspect that was the least appropriate in therapy was also moderately to highly inappropriate in all scenarios, all scenarios, except when creating a transphyseal bridge with a lag screw, except when performing castration as they exceeded the recommendation with fewer than 20% of respondents reporting timing their administration of in all instances (Fig 5). antimicrobials to generate effective serum antimicrobial concentrations at An individual’s overall optimal and suboptimal compliance was the time of surgery. In the vast majority of cases the penicillin administered calculated as the proportion of optimal or suboptimal compliant scenarios was intramuscular (i.m.) procaine penicillin, with 38% of respondents among the total number of scenarios completed in the questionnaire. administering this within 30 min before surgery and 33% administering it There was marked variation between individual respondents, ranging from after surgery. As the time required to reach maximal plasma 1% to 100% overall optimal compliance. The distribution of the proportion concentrations of penicillin after i.m. administration of procaine penicillin in of scenarios that were compliant was not normally distributed. After horses is 3.5 h [21], administration in the 30 min before surgery was adjusting for the effect of practice type, species treated, size and location, classed as inappropriate. The dose of antimicrobials was also commonly gender, position in practice, the presence or absence of post-graduate

Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd

71 Antimicrobials for surgical prophylaxis L. Y. Hardefeldt et al.

100 90 80 70 60 50 40 % respondents 30 20 10 0 Contaminated Transphyseal Eye ablation Clean wound Gelding Umbilical Dermal wound (not bridge (sutured) hernia mass sutured and not repair removal infected) Never Sometimes/rarely Always/frequently Fig 1: Frequency of antimicrobial usage for surgical prophylaxis in seven scenarios.

100

90

80

70

60

50

% antimicrobials 40

30

20

10

0 Aminopenicillins TMS* Gentamicin 3rd generation Oxytetracycline Metronidazole cephalosporin Narrow spectrum Extended spectrum Long acting Fig 2: Overall proportions of antimicrobials reported as being used in surgical prophylaxis. *TMS, Trimethoprim sulphonamide.

qualifications and presence or absence of a practice antimicrobial use commonly reported and the choice of antimicrobial drug was appropriate, policy, the odds of compliance was 3.2 times higher in university in most instances, for all scenarios in which they were indicated. The veterinarians compared with general practitioners (95% CI, 1.14–8.91). predominant use of procaine penicillin differed from a survey of There was a trend towards lower odds of compliance in recent graduates antimicrobial use by equine practitioners in the UK, in which the use of (graduated after 2011) compared with older graduates (veterinarians who trimethoprim-sulphonamide predominated [14], but this survey also graduated in or before 2011) (OR 0.60; 95% CI, 0.35–1.01; P = 0.056) and in included medical conditions and this may have affected the choice of practitioners from Western Australia (OR 0.5; 95% CI, 0.26–1.01; P = 0.056) antimicrobial. The frequent use of prophylactic antimicrobials for routine (Table 1). elective surgeries such as castration and dermal mass removal might be expected in ambulatory practice because of the need to perform surgery Discussion in exposed (outdoor) conditions. However, less than 25% of respondents selected surgical conditions as a factor important in the decision to use or This is the first survey to investigate antimicrobial use by equine not use antimicrobials. This implies that use of antimicrobials for surgical practitioners in Australia and the first to report on compliance with prophylaxis is routine in equine practice, even for clean surgeries and there antimicrobial guidelines in equine surgery. Consistent with BEVA’s is little consideration of the actual need for antimicrobial therapy in these PROTECT-ME guidelines [16], antimicrobial use for surgical prophylaxis was scenarios. There is evidence that, in some situations, clean surgical

4 Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd

72 L. Y. Hardefeldt et al. Antimicrobials for surgical prophylaxis

100

90

80

70

60

50

40

% Antimicrobials reported 30

20

10

0 Castration Clean Dermal Contaminated Umbilical Transphyseal Eye wound mass wound (not hernia bridge ablation (sutured) removal sutured and not repair infected) Penicillin TMS Pencillin/gentamicin 3rd generation cephalosporin Penicillin-long acting Other LIRA* Fig 3: Antimicrobials used for prophylaxis in each of the surgical scenarios. *LIRA, Low importance rating antimicrobial.

100 90 80 70 60 50 40

% Respondents 30 20 10 0 Castration Transphyseal Dermal Eye ablation Umbilical Clean Contaminated bridge mass hernia wound wound (not removal repair (sutured) sutured and not infected) Optimal Suboptimal Fig 4: Proportions of veterinarians reporting optimal and suboptimal compliance with BEVA guidelines for prophylactic antimicrobial use in different surgical scenarios. Suboptimal compliance reflects appropriate drug choice, but inappropriate doses or timing of antimicrobial administration to allow for adequate serum antimicrobial concentrations at the time of surgery, or a duration of therapy that was not compliant with guidelines. procedures performed without antimicrobial prophylaxis have similar The failure of guidelines to influence prescribing behaviour is multifactorial complication rates to those performed with prophylactic antimicrobial in human medicine, with a lack of appreciation of an individual’s role in therapy [22] suggesting a need for re-evaluation of the need for addressing the bigger issue [23], failure to get senior practitioners to antimicrobial therapy in every surgical case. support the use of guidelines [24,25], perceived inconvenience of Few respondents reported having an antimicrobial policy at the clinic in appropriate timing of administration of drugs [26] and interference by which they practiced (20%), although this was a higher proportion than was guidelines in clinical autonomy [27], all reported to play a role in different seen in a recent study in the UK, in which fewer than 1% of respondents scenarios. Some, or all, of these factors may also influence the successful reported having a written antimicrobial policy [14]. It was concerning that implementation of guidelines in veterinary practices. only 7% of practitioners with an antimicrobial use policy in place reported Compliance with BEVA’s PROTECT ME guidelines was used as the gold that this document was a source of information used to guide antimicrobial standard for this survey. These guidelines represent a conservative therapy. In contrast, companion animal practitioners who had an approach to antimicrobial use in surgical prophylaxis; for many other antimicrobial use policy in their practice reported using a range of different species antimicrobials are not recommended for clean surgeries and information sources to guide antimicrobial therapy compared with therapy for less than 24 h after surgery is recommended for clean- veterinarians whose practice did not have an antimicrobial use policy [18]. contaminated surgeries [28,29]. The duration of therapy recommended by

Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd

73 Antimicrobials for surgical prophylaxis L. Y. Hardefeldt et al.

100

90

80

70

60

50

40 % Respondents 30

20

10

0 Clean Umbilical Eye Dermal Transphyseal Castration wound hernia ablation mass bridge (sutured) removal

Timing inappropriate Dose inappropriate Duration inappropriate Fig 5: Proportions of veterinarians reporting suboptimal compliance with antimicrobial prophylaxis guidelines evaluated by factor. the BEVA PROTECT ME guidelines is arguably too long, but there is limited This “do not use unless” clause is legally binding, but to the authors’ evidence in the literature to guide duration of antimicrobial administration knowledge there has been no enforcement of this requirement by for surgical prophylaxis in equine medicine. There were no instances where authorities. If this situation was to change, this label requirement may drive the duration of therapy was shorter than that recommended in the BEVA veterinarians towards using antimicrobials with a higher importance rating, PROTECT-ME guidelines [16]. as alternatives for treating Gram-negative infections are limited in equine Levels of optimal compliance were very low across all scenarios in this practice. The labelling of gentamicin should be updated to allow for survey. Levels of suboptimal compliance were higher, with the appropriate empirical use in equine practice. antimicrobial commonly selected, but administration either at an The scenario of a contaminated wound over the thorax (that heals by inappropriate dose or frequency, or for too long after surgery. This secondary intention) of a horse had the lowest level of optimal or inappropriate dosing and frequency of administration of procaine penicillin suboptimal compliance in this survey. This was due to near uniform is concerning as low dose antimicrobial therapy not only promotes treatment with broad-spectrum antimicrobials, when the guidelines antimicrobial resistance, but also fails to achieve serum drug recommend no antimicrobial therapy. Such over treatment of wounds was concentrations above the minimum inhibitory concentrations for common also identified in a UK study in which 96.8% of respondents reported equine pathogens. Appropriate drug doses are taught at all Australian prescribing antimicrobial therapy for a contaminated leg wound [14]. An universities (M. Wereszka, C. Secombe, S. Raidal, R. Tan; Personal education campaign targeting antimicrobial therapy for wounds in horses communication 2016). Consistent with this, veterinarians practicing in is clearly needed. university teaching hospitals had higher odds of optimal compliance with There was a statistically insignificant trend towards lower optimal guidelines compared with general practitioners. The labelling of procaine compliance with guidelines for recentgraduatescomparedwitholder penicillin by manufacturers in Australia is misleading (labelled dose 20 ml/ graduates. Recent graduates were also found to have lower optimal 500 kg i.m. daily [30], with the doses recommended on the labels falling compliance in a survey of prophylactic antimicrobial therapy in companion below those now recognised as appropriate (22,000 iu/kg bwt every 12 h) animal surgery [18]. It is possible this is due to poor university teaching, and this may contribute to the problem of underdosing. The labelling for but it seems more likely that lack of confidence in recently graduated gentamicin is also misleading, with the labelled dose rate (2 mg/kg bwt veterinarians results in their overuse of antimicrobials for surgical three times daily) differing considerably from the current recommended prophylaxis. Inappropriate labelling may also contribute, as recently dose rates (6.6–10 mg/kg bwt once daily). Use of the dose rates suggested graduated veterinarians may rely more heavily on the label as a source of by the label is likely to significantly increase the risk of acute renal failure in information than more experienced practitioners. treated horses. However, the doses for gentamicin therapy actually There are several features of this study that may have influenced the reported by respondents were largely appropriate (89% of dose rates results. Recall bias is a common problem with questionnaire-based surveys appropriate), whereas those for procaine penicillin were largely where study participants are asked to recall past events. Hypothetical inappropriate (32% of dose rates appropriate). The label is clearly not the scenarios were posed rather than asking clinicians to recall specific cases only factor leading to use of inappropriate doses of penicillin by to minimise this. Respondents were self-selected in this study and many veterinarians. Changes to legislation are needed to ensure that were recruited at conferences, so practitioners who were more likely to antimicrobial drug labels are regularly updated to reflect the dose needed complete continuing education and had more awareness of recommended to effectively and safely treat common equine pathogens. Long-acting prescribing practices, may be over-represented. Veterinarians from the penicillin represented 3% of antimicrobials used for surgical prophylaxis in states in which the conferences were held may also be over-represented, this survey. This formulation of penicillin is inappropriate in equine but the results were adjusted to correct for this lack of equivalence in medicine as it fails to generate plasma concentrations of penicillin above sampling. Other adjustments to correct for population differences were not minimum inhibitory concentrations for common equine pathogens [31]. possible because there are no population statistics for the Australian Practitioners commonly reported that they used gentamicin for surgical veterinary profession. However, all age groups were represented, as were prophylaxis in this survey (12% of antimicrobials reported). In Australia, the rural and urban veterinarians. The survey was anonymous, to minimise current label dictates that gentamicin only be used after culture and response bias. sensitivity testing that indicates that it is the only appropriate antimicrobial, In conclusion, this survey has shown that antimicrobials are commonly a requirement that clearly cannot be met for use in surgical prophylaxis. used for surgical prophylaxis in equine practice in Australia and that

6 Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd

74 L. Y. Hardefeldt et al. Antimicrobials for surgical prophylaxis appropriate antimicrobial agents are generally chosen, with procaine 5. Weese, J.S., Dick, H., Willey, B.M., McGeer, A., Kreiswirth, B.N., Innis, B. penicillin the most commonly used drug and antimicrobials of high and Low, D.E. (2006) Suspected transmission of methicillin-resistant importance rating rarely used. Education is warranted to improve drug Staphylococcus aureus between domestic pets and humans in veterinary dosing, timing of administration prior to surgery and to shorten the clinics and in the household. Vet. Microbiol. 115, 148-155. duration of surgical prophylaxis. Legislation should also be amended to 6. Platell, J.L., Cobbold, R.N., Johnson, J.R., Heisig, A., Heisig, P., Clabots, C., require that product labels carry appropriate current dosing advice. In Kuskowski, M.A. and Trott, D.J. (2011) Commonality among addition, the use of antimicrobials for uncomplicated wounds should be fluoroquinolone-resistant sequence type ST131 extraintestinal Escherichia discouraged much more strongly across the veterinary profession. coli isolates from humans and companion animals in Australia. Antimicrob. Agents Chemother. 55, 3782-3787. 7. Liu, W., Liu, Z., Yao, Z., Fan, Y., Ye, X. and Chen, S. (2015) The prevalence Authors’ declaration of interests and influencing factors of methicillin-resistant Staphylococcus aureus carriage in people in contact with livestock: a systematic review. Am. J. No competing interests have been declared. Infect. Control 43, 469-475. 8. Bosch, T., Verkade, E., van Luit, M., Landman, F., Kluytmans, J. and Schouls, L.M. (2015) Transmission and persistence of livestock-associated Ethical animal research methicillin-resistant Staphylococcus aureus among veterinarians and their household members. Appl. Environ. Microbiol. 81, 124-129. This research was approved by the University of Melbourne Faculty of 9. Ishihara, K., Shimokubo, N., Sakagami, A., Ueno, H., Muramatsu, Y., Veterinary and Agricultural Sciences Human Ethics Advisory Group under Kadosawa, T., Yanagisawa, C., Hanaki, H., Nakajima, C., Suzuki, Y. and Approval No. 1646102. Completion of the questionnaire was taken as Tamura, Y. (2010) Occurrence and molecular characteristics of methicillin- participant consent. resistant Staphylococcus aureus and methicillin-resistant Staphylococcus pseudintermedius in an academic veterinary hospital. Appl. Environ. Microbiol. 76, 5165-5174. Source of funding 10. Walther, B., Hermes, J., Cuny, C., Wieler, L.H., Vincze, S., Abou Elnaga, Y., Stamm, I., Kopp, P.A., Kohn, B., Witte, W., Jansen, A., Conraths, F.J., This research was funded by the National Health and Medical Research Semmler, T., Eckmanns, T. and Lubke-Becker, A. (2012) Sharing more Council through the Centres of Research Excellence programme, grant than friendship-nasal colonization with coagulase-positive staphylococci number 1079625. L. Hardefeldt was a recipient of an Australian Post- (CPS) and co-habitation aspects of dogs and their owners. PLoS ONE 7, graduate Award scholarship. e35197. 11. Anon. (2014) Quantity of antimicrobial products sold for veterinary use in Australia: July 2005 to June 2010, Australian Pesticides and Veterinary Acknowledgements Medicines Authority http://www.apvma.gov.au/. Thanks to the Australian Veterinary Association, Equine Veterinarians 12. Weese, J.S. and Sabino, C. (2005) Scrutiny of antimicrobial use in racing Australia and the state veterinary boards for their support in the delivery of horses with allergic small airway inflammatory diseases. Can. Vet. J. 46, the survey. 438-439. 13. Weese, J.S. and Cruz, A. (2009) Retrospective study of perioperative antimicrobial use practices in horses undergoing elective arthroscopic Authorship surgery at a veterinary teaching hospital. Can. Vet. J. 50, 185-188. 14. Hughes, L.A., Pinchbeck, G., Callaby, R., Dawson, S., Clegg, P. and L. Hardefeldt, G. Browning, K Thursky, J. Gilkerson, H. Billman-Jacobe and K. Williams, N. (2013) Antimicrobial prescribing practice in UK equine Bailey were involved in the study design. L. Hardefeldt, G. Browning, J. veterinary practice. Equine Vet. J. 45, 141-147. Gilkerson and K. Bailey were involved in study execution. L. Hardefeldt and 15. Anon. (2015) Importance rating and summary of antibacterials used in M. Stevenson were involved in the data analysis and interpretation. L. human health in Australia, Commonweath of Australia, Australian Hardefeldt was the primary author of the manuscript, with the assistance Strategic and Technical Advisory Group on Antimicrobial Resistance http:// of K. Bailey and G. Browning. All authors reviewed the manuscript and www.health.gov.au/internet/main/publishing.nsf/Content/ohp-amr.htm. gave it their final approval. 16. Anon. (2016) PROTECT ME, British Equine Veterinary Association. http:// www.beva.org.uk/useful-info/Vets/Guidance/AMR. 17. Nelson, R.L., Glenny, A.M. and Song, F. (2009) Antimicrobial prophylaxis Manufacturer’s address for colorectal surgery. Cochrane Database Syst. Rev. 21, CD001181. 18. Hardefeldt, L.Y., Browning, G.F., Thursky, K., Gilkerson, J.R., Billman- aMicrosoft Corporation, Redmond, WA, USA Jacobe, H., Stevenson, M.A. and Bailey, K.E. (2017) Antimicrobials used for surgical prophylaxis by companion animal veterinarians in Australia. Vet. Microbiol. 203, 301-307. References 19. Anon. (2015) Report on Projection Modelling for the Veterinarian Workforce, Australian Veterinary Association, THINC http://www.ava.com.au/. 1. Jiang, X., Yang, H., Dettman, B. and Doyle, M.P. (2006) Analysis of fecal 20. Harris, P.A., Taylor, R., Thielke, R., Payne, J., Gonzalez, N. and Conde, J.G. microbial flora for antibiotic resistance in ceftiofur-treated calves. (2009) Research electronic data capture (REDCap) – A metadata-driven Foodborne Pathog. Dis. 3, 355-365. methodology and workflow process for providing translational research 2. Rentala, M., Lahti, E., Kuhalampi, J., Pesonen, S., Jarvinen, A.K., Saijonmaa- informatics support. J. Biomed. Inform. 42, 377-381. Koulumies, L. and Honkanen-Buzalski, T. (2004) Antimicrobial resistance in 21. Ubon, C.E., Soma, L.R., Luo, Y., McNamara, E., Fennell, M.A., May, L., Staphlococcus spp., Escherichia coli and Enterococcus spp. in dogs given Teleis, D.C., Rudy, J.A. and Watson, A.O. (2000) Pharmacokinetics of antibiotics for chronic dermatological disorders compared with non- penicillin G procaine versus penicillin G potassium and procaine treated control dogs. Acta Vet. Scand. 45, 37-45. hydrochloride in horses. Am. J. Vet. Res. 61, 811-815. 3. Leite-Martins, L.R., Mahu, M.I., Costa, A.L., Mendes, A., Lopes, E., 22. Borg, H. and Carmalt, J.L. (2013) Postoperative septic arthritis after Mendonca, D.M., Niza-Ribeiro, J.J., de Matos, A.J. and da Costa, P.M. elective equine arthroscopy without antimicrobial prophylaxis. Vet. Surg. (2014) Prevalence of antimicrobial resistance in enteric Escherichia coli 42, 262-266. from domestic pets and assessment of associated risk markers using a 23. Giblin, T.B., Sinkowitz-Cochran, R.L., Harris, P.L., Jacobs, S., Liberatore, K., generalized linear mixed model. Prev. Vet. Med. 117, 28-39. Palfreyman, M.A., Harrison, E.I., Cardo, D.M.; CDC Campaign to Prevent 4. Anon. (2014) Antimicrobial Resistance: Tackling a crisis for the health and Antimicrobial Resistance Team. (2004) Clinicians’ perceptions of the wealth of nations, Ed: J. O’Neill, The Review on Antimicrobial Resistance problem of antimicrobial resistance in health care facilities. Arch. Intern. http://www.amr-review.org/. Med. 164, 1662-1668. 7 Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd 5 Antimicrobials for surgical prophylaxis L. Y. Hardefeldt et al.

24. De Souza, V., MacFarlane, A., Murphy, A.W., Hanahoe, B., Barber, A. and 29. Anon. (2013) Clinical practice guidelines for antimicrobial prophylaxis in Cormican, M. (2006) A qualitative study of factors influencing surgery, US Department of Health and Human Services. http://www. antimicrobial prescribing by non-consultant hospital doctors. J. guideline.gov/. Antimicrob. Chemother. 58, 840-843. 30. MSD Animal Health. (http://www.infopest.com.au/) Depocillin Procaine 25. Cortoos, P.J., De Witte, K., Peetermans, W.E., Simoens, S. and Laekeman, Penicillin 300 ng/ml Injection, Growcom. G. (2008) Opposing expectations and suboptimal use of a local antibiotic 31. Love, D.N., Rose, R.J., Martin, I.C. and Bailey, M. (1983) Serum hospital guideline: a qualitative study. J. Antimicrob. Chemother. 62, 189- concentrations of penicillin in the horse after administration of a variety 195. of penicillin preparations. Equine Vet. J. 15, 43-48. 26. Tan, J.A., Naik, V.N. and Lingard, L. (2006) Exploring obstacles to proper timing of prophylactic antibiotics for surgical site infections. Qual. Saf. Health Care 15, 32-38. Supporting Information 27. Cribb, A. and Barber, N. (1997) Prescribers, patients and policy: the limits of technique. Health Care Anal. 5, 292-298. Additional Supporting Information may be found in the online version 28. Anon. (2016) PROTECT, British Small Animal Veterinary Association http:// of this article at the publisher’s website: www.bsava.com/Resources/PROTECT.aspx. Supplementary Item 1: Copy of questionnaire used in the survey.

7

8 Equine Veterinary Journal 0 (2017) 1–8 © 2017 EVJ Ltd 6 BACK TO TABLE OF CONTENTS

Chapter 6:

DETAILED ANTIMICROBIAL USE BY BOVINE VETERINARIANS

7

7 10.1136/vr.104375

Paper Paper Cross-sectional study of antimicrobials used for surgical prophylaxis Veterinary Record (2017) doi: 10.1136/vr.104375 Laura Y Hardefeldt, BSc, BVMS, for Antimicrobial Stewardship, Peter by bovine veterinary DACVIM, Doherty Institute, Carlton, Victoria, Glenn F Browning, BVSc, PhD, Australia James R Gilkerson, BVSc, PhD, Laura Y Hardefeldt is also at National practitioners in Australia Helen Billman-Jacobe, BSc, PhD, Centre for Antimicrobial Stewardship, Mark A Stevenson, BVSc, PhD, Peter Doherty Institute, Carlton, Victoria, Kirsten E Bailey, BVSc, Australia Laura Y Hardefeldt, Glenn F Browning, Faculty of Veterinary and Agricultural Karin A Thursky, James R Gilkerson, Sciences, Asia-Pacific Centre for Glenn F Browning is also at National Animal Health, University of Melbourne, Centre for Antimicrobial Stewardship, Helen Billman-Jacobe, Mark A Stevenson, Parkville, Victoria, Australia Peter Doherty Institute, Carlton, Victoria, Kirsten E Bailey Karin A Thursky, MBBS, Australia National Centre for Antimicrobial Karin A Thursky is also at National Stewardship, Peter Doherty Institute, Centre for Antimicrobial Stewardship, Antimicrobials are widely used in veterinary practices, but Carlton, Victoria, Australia Peter Doherty Institute, Carlton, Victoria, there has been no investigation of antimicrobial classes Australia Laura Y Hardefeldt is also at National used or the appropriateness of their use in bovine practice. Centre for Antimicrobial Stewardship, Helen Billman-Jacobe is also at National This study investigated antimicrobial use for surgical Peter Doherty Institute, Carlton, Victoria, Centre for Antimicrobial Stewardship, prophylaxis in bovine practice in Australia. A cross-sectional Australia Peter Doherty Institute, Carlton, Victoria, study of veterinarian antimicrobial usage patterns was Australia Glenn F Browning is also at National conducted using an online questionnaire. Information Centre for Antimicrobial Stewardship, Kirsten E Bailey is also at National Centre solicited included respondent’s details, the frequency Peter Doherty Institute, Carlton, Victoria, for Antimicrobial Stewardship, Peter with which antimicrobials were used for specific surgical Australia Doherty Institute, Carlton, Victoria, conditions (including the dose, timing and duration of Australia Karin A Thursky is also at National therapy) and details of practice antimicrobial use policies Centre for Antimicrobial Stewardship, Laura Y Hardefeldt is also at National and sources of information about antimicrobials. In total, Peter Doherty Institute, Carlton, Victoria, Centre for Antimicrobial Stewardship, 212 members of the Australian veterinary profession Australia Peter Doherty Institute, Carlton, Victoria, working in bovine practice completed the survey. Australia Helen Billman-Jacobe is also at National Antimicrobials were always or frequently used by more Centre for Antimicrobial Stewardship, Glenn F Browning is also at National than 75 per cent of respondents in all scenarios. Generally, Peter Doherty Institute, Carlton, Victoria, Centre for Antimicrobial Stewardship, antimicrobial drug choice was appropriate for the reported Australia Peter Doherty Institute, Carlton, Victoria, surgical conditions. Procaine penicillin and oxytetracycline Australia Kirsten E Bailey is also at National Centre accounted for 93 per cent of use. However, there was a wide for Antimicrobial Stewardship, Peter Karin A Thursky is also at National range of doses used, with underdosing and inappropriate Doherty Institute, Carlton, Victoria, Centre for Antimicrobial Stewardship, timing of administration being common reasons for Australia Peter Doherty Institute, Carlton, Victoria, inappropriate prophylactic treatment. There was very low Australia Laura Y Hardefeldt is also at National use of critically important antimicrobials (3.3 per cent of Centre for Antimicrobial Stewardship, Helen Billman-Jacobe is also at National antimicrobials reported). Antimicrobial use guidelines need Peter Doherty Institute, Carlton, Victoria, Centre for Antimicrobial Stewardship, to be developed and promoted to improve the responsible Australia Peter Doherty Institute, Carlton, Victoria, use of antimicrobials in bovine practice. Glenn F Browning is also at National Australia Centre for Antimicrobial Stewardship, Kirsten E Bailey is also at National Centre Introduction Peter Doherty Institute, Carlton, Victoria, for Antimicrobial Stewardship, Peter Antimicrobial use in humans and animals generates selective Australia Doherty Institute, Carlton, Victoria, pressure that increases the prevalence of antimicrobial resistance 1–3 Karin A Thursky is also at National Australia in bacterial populations. With the growing threat of antimi- crobial resistant bacteria in medical hospitals, the community Centre for Antimicrobial Stewardship, E-mail for correspondence: Peter Doherty Institute, Carlton, Victoria, and in animals, there is an increasing focus on veterinary anti- laura.hardefeldt@ unimelb. edu. au 4 Australia microbial usage and many global antimicrobial resistance strat- Provenance and peer review Not egies emphasise antimicrobial stewardship in both human and Helen Billman-Jacobe is also at National commissioned; externally peer veterinary medicine.5–7 In addition, veterinarians, farm workers Centre for Antimicrobial Stewardship, reviewed. and their families have been shown to have a risk of acquiring Peter Doherty Institute, Carlton, Victoria, multidrug-resistant bacteria from livestock.8–10 Australia Received April 14, 2017 Data on quantities of veterinary antimicrobials in Australia Revised July 25, 2017 Kirsten E Bailey is also at National Centre are limited to periodic reports by the Australian Pesticides and Accepted August 4, 2017

10.1136/vr.104375 | Veterinary Record | 1 of 6 7

8 Paper Paper

Veterinary Medicines Authority, which records total volumes of TABLE 1: Survey respondent demographics. antimicrobials imported for use in the veterinary and agricultural sectors.11 However, the distribution of use of these antimicro- Australian veterinary bials cannot be tracked further except for specific formulations, Survey respondents workforce* Characteristic N (%) % such as intramammary therapies, where use is largely limited to the treatment of clinical and subclinical mastitis in dairy cows. Gender Recent publications from Europe have investigated the common Male 89 (42) 39 reasons for antimicrobial use in cattle12 and factors that influence Female 123 (58) 61 prescribing habits among veterinarians.13 14 However, there has Location NA been no investigation of antimicrobials used for surgical proph- Metropolitan 17 (8) ylaxis in cattle. In addition, the classes of antimicrobial used, the Rural 195 (91) appropriateness of the doses administered and the duration of Age (years) NA therapy in cattle have not been evaluated. 20–30 99 (46) The Australian Strategic and Technical Advisory Group on 31–40 53 (25) Antimicrobial Resistance (ASTAG) issued an importance rating 41–50 30 (14) and summary of antibacterials used in human health in Australia 51–60 16 (7.5) 15 in 2015. Those given a high importance rating include pipera- >60 14 (6.6) cillin-tazobactam, ticarcillin-clavulanate, the third-generation State and fourth-generation cephalosporins, aztreonam, tigecycline, ACT 1 (0.5) 3.0 vancomycin, teicoplanin, amikacin, the streptogramins, fluoro- NSW 48 (23) 28 quinolones and rifampicin. The ASTAG have recommended that NT 3 (1.4) 0.8 these antimicrobials should be used as third-line therapies, that is, SA 13 (5.7) 6.0 they should only be used when culture and susceptibility testing QLD 39 (18) 24 or other compelling clinical evidence justifies their use. The TAS 7 (3.3) 2.2 third-generation cephalosporins are the only critically important VIC 91 (43) 25 antimicrobials that are registered for use in cattle in Australia. WA 11 (5.2) 11 There are no guidelines for antimicrobial use in cattle in Australia. Despite this, appropriateness of drug doses and timing Number of veterinarians in practice NA of administration can be deduced from the pharmacokinetics of 1 or 2 35 (16) antimicrobials in cattle. Guidelines provide a necessary first step 3 or 4 65 (31) towards the implementation of veterinary antimicrobial steward- 5–10 84 (40) ship, but audit and feedback are necessary to improve prescribing 11–20 24 (11) practices. Antimicrobial use for surgical prophylaxis has been >20 4 (1.9) an area in human medicine where application of guidelines and *Australian Veterinary Association.51 monitoring has led to more appropriate antimicrobial therapy.16 NA, not available. The aim of this study was to investigate self-reported antimi- crobial use in a range of surgical conditions in bovine practice in This research was approved by the University of Melbourne Australia and to assess appropriateness of drug doses and timing Faculty of Veterinary and Agricultural Sciences Human Ethics of administration to prevent surgical site infection. Advisory Group under Approval No. 1646102. Materials and methods The survey study population and distribution is described else- Results where.17 The study population comprised those veterinarians A total of 212 members of the Australian veterinary profes- who completed surgery on cattle as part of their weekly practice sion completed the section of the survey addressing bovine (estimated to be 240018 in 2015). Sample size calculations were practice. All states and territories were represented, as were carried out to determine the number of respondents required to recent and older graduates (Table 1). Livestock-only practi- make appropriate inferences from the survey. To be 95 per cent tioners represented only 8 per cent (95 per cent CI 4 per cent certain that our estimate of the population prevalence of veter- to 12 per cent) of respondents, while the remaining 92 per cent inarians using a given class of antimicrobials was within 7.5 per (95 per cent CI 88 per cent to 96 per cent) treated a mixture of cent of the true population prevalence of 50 per cent, a total of species (4.2 per cent bovine and equine, 6.6 per cent bovine and 160 completed surveys were required. Sample size calculations companion animals, 81 per cent bovine, equine and companion were carried out assuming a 50 per cent population prevalence animals) and 23 per cent of the respondents treated only dairy because this provided the largest sample size estimate for a cattle, 45 per cent treated only beef cattle and 32 per cent treated constant margin of error. both dairy and beef cattle. A relatively small proportion of prac- Survey details are described elsewhere.17 The bovine surgical tices had an antimicrobial use policy (22 per cent, 95 per cent CI scenarios included in the survey were correction of a left displaced 16 per cent to 28 per cent). A wide range of sources of informa- abomasum, caesarean section, eye ablation, exploratory lapa- tion on antimicrobials for surgical prophylaxis were used, with rotomy and repair of an umbilical hernia. no single source of information predominating. Practitioners Data were downloaded from the survey software to spread- reported using experience (56 per cent), textbooks (52 per cent) sheets (Microsoft Office Excel, 2016). The entire bovine section of and undergraduate course notes (51 per cent) most frequently, the survey had to be completed by the respondent to be included with continuing education (50 per cent), colleagues (48 per cent) in the analysis. Descriptive statistics were computed with and the product label (43 per cent) also frequently cited. Having percentages being reported as the proportion of the total respond- an antimicrobial use policy did influence the sources of informa- ents answering a given question. Where respondents reported tion used by some veterinarians, but only 38 per cent of practi- that they did not perform a specific surgery, these individuals tioners who reported having an antimicrobial use policy in place were excluded from the analyses of that question. Comparison of cited this practice policy as a source of information. The amount proportions was performed using 2 tests using functions within of contamination during surgery (82 per cent) and surgical condi- Stata V.13. tions (60 per cent) were the most frequently reported influences

2 of 6 | Veterinary Record | 10.1136/vr.104375 7

9 Paper

100

90

80

70

60

50

40 % respondents 30

20

10

0 Caesarian section Left displaced Exploratory Repair of umbilical Eye ablation abomasum laparotomy hernia

Never Sometimes/Rarely Always/Frequently FIG 1: Frequency of prophylactic antimicrobial usage in cattle for different surgical scenarios on the decision to prescribe antimicrobials in the surgical Overall, the most frequently prescribed antimicrobial classes scenarios. in this survey were penicillins (47 per cent) and tetracyclines The five categories indicating the frequency of antimicrobial (specifically oxytetracycline, 46 per cent). All other classes repre- use for each surgical condition were combined into three groups sented less than 7 per cent of all reported antimicrobials in the (always/frequently, sometimes/rarely and never). The majority survey (third-generation cephalosporins 3.3 per cent, trimeth- of respondents indicated that they administered antimicrobials oprim sulphonamides 1.9 per cent, erythromycin 1.1 per cent always or frequently in all scenarios. Eye ablation and repair of and tulathromycin 0.2 per cent) (fig 2). There were no significant an umbilical hernia were the only scenarios in which fewer than differences in the classes of antimicrobials used in the different 80 per cent of respondents indicated that they administered anti- surgical scenarios (fig 3). There was a very low prevalence of microbials always or frequently. Eye ablation had the least anti- use of critically important antimicrobials (3.3 per cent), with microbial use; antimicrobials were used always or frequently by third-generation cephalosporins the only drugs with this rating 76 per cent of respondents, sometimes or rarely by 13 per cent of reported, and use not exceeding 6 per cent of antimicrobials used respondents and never by 11 per cent of respondents. The pattern in any one scenario (fig 3). There was wide variation in duration of response was similar for surgery to repair an umbilical hernia— of therapy across scenarios. Eye ablation and repair of an umbil- antimicrobials were used always or frequently by 77 per cent of ical hernia were the only scenarios in which antimicrobial therapy respondents, sometimes or rarely by 13 per cent of respondents was halted within 24 hours by more than 40 per cent of respond- and never by 10 per cent of respondents (fig 1). ents (45 per cent and 42 per cent, respectively). Antimicrobial

50

45

40

35

30

25

20 % antimicrobials

15

10

5

0 Procaine Oxytetracycline 3rd generation Erythromycin TMS Extended Tulathromycin penicillin cephalosporin spectrum penicillins

Short acting Long acting Pessary/ocular formulation FIG 2: Overall proportion of antimicrobials used for surgical prophylaxis across all scenarios. TMS, trimethoprim sulphonamide

10.1136/vr.104375 | Veterinary Record | 3 of 6

80 Paper Paper

100 intraperitoneal or incisional antimicrobials during surgery. There was no significant difference in the proportions of recent gradu- 90 ates (graduated after 2010) and older graduates using appropriate 80 doses (70 per cent and 67 per cent, respectively, p=0.97) nor in the proportion of these groups appropriately timed antimicrobial 70 administration (48 per cent and 44 per cent, respectively, p=0.96). 60 Similarly, there was no significant difference in proportions of male and female respondents using appropriate doses (32 per cent 50 and 33 per cent respectively, p=0.99) nor in the proportions of 40

% respondents these groups using appropriately timed antimicrobial administra- tion (47 per cent and 45 per cent, respectively, p=0.98). There was 30 also no significant difference between respondents from small 20 practices (one or two veterinarians) and respondents from larger practices (more than two veterinarians) in either the proportions 10 using appropriate doses (17 per cent and 16 per cent, respectively, 0 p=0.99) nor in the proportions using appropriately timed antimi- Left displaced Caesarean Eye ablation Exploratory Repair of crobial administration (31 per cent and 48 per cent, respectively, abomasum section laparotomy umbilical hernia p=0.89).

Penicillins Oxytetracycine 3rd generation cephalosporin LIRA Discussion FIG 3: Proportions of different classes of antimicrobials for To the best of our knowledge, this is the first survey to investi- surgical prophylaxis in specific scenarios. LIRA, other low gate antimicrobial use by bovine veterinarians in Australia. We importance rating antimicrobials found that antimicrobials are used by most veterinarians for all surgical scenarios presented and that -lactams (predominately therapy was longest for correction of a left displaced abomasum, procaine penicillin) and oxytetracycline were the antimicrobials with 49 per cent of respondents indicating therapy was typically that were most commonly administered. The frequent use of continued for four to seven days (fig 4). prophylactic antimicrobials for routine elective surgeries, such The dose of procaine penicillin and oxytetracycline adminis- as hernia repairs, might be expected in ambulatory practice due tered for surgical prophylaxis varied widely. The dose of procaine to the necessity to perform surgery in exposed (outdoor) condi- penicillin administered ranged from 7.5 to 24 mg/kg, with the tions. Consistent with this, 60 per cent of respondents indicated most frequent dose rates being 12.5 mg/kg (37 per cent of respond- that surgical conditions influenced their decision making about ents) and 15 mg/kg (43 per cent of respondents). Procaine peni- antimicrobial therapy for surgical prophylaxis. In addition, the cillin was administered once daily by 96 per cent of respondents amount of contamination was the most cited factor influencing using a short acting formulation, with less than 5 per cent of antimicrobial use (82 per cent of respondents). The duration of respondents reporting 12 hourly administration. Similarly, for therapy was longest for correction of a left displaced abomasum, oxytetracycline the dose range was wide (2.5–13.5 mg/kg), with which is not an emergency procedure and does not require enter- the most frequent dose rates being 4–5 mg/kg (52 per cent of otomy, so should be able to be performed as a clean procedure. respondents) and 10 mg/kg (38 per cent of respondents). Most This suggests that use of antimicrobials for surgical prophylaxis respondents reported administering oxytetracycline once daily is routine in bovine practice in Australia, even for clean surgeries, when using a short-acting formulation (97 per cent), with the and that there is little consideration of the need for antimicrobial remainder reporting that they administered it every 12 hours therapy in these scenarios. There is evidence that, in some situa- (2.8 per cent). Half of the respondents reported that they admin- tions, clean surgical procedures performed without antimicrobial istered antimicrobials for surgical prophylaxis before surgery prophylaxis have similar complication rates to those with anti- and 50 per cent indicated that they administered them after microbial therapy,19 which should prompt a re-evaluation of the surgery. In addition, 20 per cent of respondents indicated using need for antimicrobial therapy in every surgical case. A minority of respondents reported having an antimicrobial 100 use policy in the clinic at which they practised (22 per cent). However, as more than 90 per cent of the veterinarians who 90 completed the survey were working in mixed species veteri- 80 nary clinics, it is not clear whether these policies were only for companion animal species, or for all species treated by the practice. 70 Consistent with studies on European veterinarians, experience,

60 published literature and course notes were the most cited sources for information about antimicrobial use.13 Experience was also the 50 most important basis for decision making about antimicrobial use by cattle veterinarians in Ireland.14 In contrast to equine veteri- 40 narians in Australia,20 but similar to companion animal veterinar- 17 30 ians, the presence of an antimicrobial use policy in the practice % respondents did change the sources of information used for decisions about 20 use of antimicrobials for surgical prophylaxis, with 38 per cent 10 of veterinarians working in a practice with a policy citing this as an important source of information compared with 13 per cent 0 of those working in a practice without such a policy. However, Eye ablation Repair of Caesarean Exploratory Left umbilical section laparotomy displaced more than 40 per cent of practitioners with an antimicrobial use hernia abomasum policy in their practice did not identify this as an important source ≤24h 1-3 days 4-7 days >7 days of information. The failure of guidelines to influence antimicro- FIG 4: Proportions of respondents indicating differing durations of bial prescribing behaviour in human medicine is multifactorial, antimicrobial prophylactic therapy for specific surgical scenarios with a lack of appreciation of an individual’s role in addressing the

4 of 6 | Veterinary Record | 10.1136/vr.104375

81 Paper wider issue of antimicrobial resistance,21 the failure to get senior evidence-based dosage recommendations, nor are data describing practitioners to support the use of guidelines,22 23 the perceived the organisms commonly responsible for surgical site infections inconvenience of appropriate timing of administration of drugs24 in cattle. Skin pathogens such as Streptococcus and Staphylococcus and interference of guidelines in clinical autonomy25 having all species, enteric pathogens such as Escherichia coli and Enterobacter been found to play a role in different scenarios. Some, or all, of species and environmental pathogens could all be expected to play these factors may also influence the successful implementation of a role in surgical site infections in cattle. Levels of resistance in guidelines into bovine veterinary practice. isolates of E coli from cattle in Australia are very low42 43 as they Antimicrobials are used for surgical prophylaxis to reduce the are in many parts of Europe,44 but data on other pathogens are risk of surgical site infections.26–28 For this to be effective, there lacking. must be adequate serum levels of antimicrobials present at the Evidence from human medicine suggests that surgical antimi- time of surgery,29 which is reliant on administration of an effective crobial prophylaxis for longer than 24 hours provides no benefit dose of antimicrobial at the appropriate time. Appropriate timing compared with administration for less than 24 hours.45 Similarly, of administration can be estimated from pharmacokinetic prop- in bovine surgery there has been no difference detected between erties of antimicrobials (time to maximal antimicrobial concen- single-dose prophylaxis and seven-day postoperative therapy tration) as can the repeat dosing intervals (twice the elimination following rumenotomy46 or correction of a caecal torsion.47 In half-life). For procaine penicillin and oxytetracycline administered addition, there were no surgical site infection reported in a series intramuscularly appropriate timing would require administration of uncomplicated hernia repairs even though 30 per cent of these 48 two30 and eight31 hours, respectively, before surgical incision. cases did not receive any prophylactic antimicrobial therapy. 49 50 If oxytetracycline is administered intravenously, administra- Antimicrobial use guidelines for companion animals and equids tion 30 minutes before surgery should allow for adequate tissue also recommend discontinuing therapy within 24 hours. Clearly, concentrations at the surgical site. In this survey, around half despite the challenging environmental conditions, cattle surgery of the veterinarians reported administering antimicrobials after can be performed in a way that minimises the need for extended surgery was performed, and thus clearly not achieving this crit- courses of prophylactic antimicrobials. In this survey, most ical goal. In addition, the doses of procaine penicillin used varied, respondents suggested that they used antimicrobial therapy for and were predominately lower than those recommended, and the longer than 24 hours in all surgical scenarios, with one to three frequency of administration was generally once daily. This dose days and four to seven days of therapy being most frequently and frequency of administration for procaine penicillin is unlikely selected in all but two scenarios posed (eye ablation and repair of to generate serum levels effective against common bovine path- umbilical hernia, where therapy for less than 24 hours was most ogens.32 Administration of subtherapeutic doses may promote frequently selected). Although relatively few practitioners indi- the development of antimicrobial resistance. Research is needed cated that they used antimicrobials for longer than seven days for to establish minimum inhibitory concentrations for common uncomplicated eye ablation, umbilical hernia repair and caesarean bovine pathogens to allow for evidence-based recommendations section, such usage is concerning, as it is likely to be both excessive and unnecessary. on treatment regimes. This is especially relevant in bovine prac- There are several features of this study that may have influ- tice, as off-label administration, at a higher dose or more frequent enced the results. Recall bias may occur with questionnaire-based administration, has implications for withholding periods for meat surveys when respondents are asked to remember events that and milk. The lack of data for withholding periods for meat and have occurred in the past. Hypothetical scenarios were posed milk for appropriate doses of antimicrobials is probably a signifi- rather than asking clinicians to recall specific cases in order to cant barrier to antimicrobial stewardship in cattle practice. minimise this. Respondents were self-selected in this study and The use of intraperitoneal and intraincisional antimicrobials many were recruited at conferences, so selection bias may also be by 20 per cent of the respondents to this survey is lower than in a present. This may have biased the results towards practitioners similar survey of Canadian veterinarians, in which more than half who were more likely to complete continuing education, and who of the respondents reported using intraoperative antimicrobials.33 had more awareness of recommended prescribing practices. The The intraperitoneal route is used in human abdominal surgery, survey was anonymous to minimise response bias. but the recommended formulations are those suitable for intrave- In conclusion, this survey has shown that, while antimicro- nous use and these are diluted in lavage solutions.34 The intraperi- bials were commonly used for surgical prophylaxis in bovine toneal use of some intramuscular formulations has led to severe practice in Australia, the choice of antimicrobial agent was gener- 35 but intra-abdominal inflammation in cattle in some instances, ally appropriate for the surgical scenario, with mainly procaine safety has not been established in most cases. Research supporting penicillin and oxytetracycline administered and very little use of efficacy is lacking in both veterinary and human medicine. critically important antimicrobials. Education is warranted to Similarly, for intraincisional antimicrobials, in human surgery improve drug dosing and timing of administration before surgery. some efficacy has been reported with infiltration of aqueous Further investigation into the appropriate duration of antimicro- 36 37 solutions before surgical incision, but there are no reports of bial therapy to prevent surgical site infections in cattle is needed. studies in veterinary medicine. Formulations used by respondents Finally, further research into minimum inhibitory concentrations to this survey were predominately intramammary preparations of common bovine pathogens and a subsequent review of drug (for intraincisional applications) and procaine penicillin (for intra- labelling is needed to ensure that suppliers are encouraged to peritoneal applications), for which there is no evidence for safety revise their labels to reflect the current understanding of antimi- nor efficacy (data not shown). crobial pharmacokinetics and pharmacodynamics and to ensure As with procaine penicillin, reported doses of oxytetracy- advice about withholding periods for meat and milk is accurate. cline varied widely. Serum concentrations of oxytetracycline can be maintained above 1 µg/mL for 24 hours after intramuscular Correction notice This article has been corrected since it published Online First. injection at a dose of 5 mg/kg,38 which may be appropriate for The author name ’Glen’ was corrected to ’Glenn’. 39 40 41 some bovine pathogens, but not for others. Administration Funding National Health and Medical Research Council through the Centres of of oxytetracycline at 10 mg/kg is much more likely to main- Research Excellence programme, grant no: 1079625. tain serum levels above minimum inhibitory concentrations Competing interests None declared. for bovine pathogens over a 24-hour period.31 Importantly, no © British Veterinary Association (unless otherwise stated in the text of the article) current minimum inhibitory concentration data on Australian 2017. All rights reserved. No commercial use is permitted unless otherwise expressly bovine pathogens are readily available to enable development of granted.

10.1136/vr.104375 | Veterinary Record | 5 of 6

82 Paper Paper

References 27 RONALD AR. Antimicrobial prophylaxis in surgery. Surgery 1983;93:172–3. 1 JIANG X, YANG H, DETTMAN B, et al. Analysis of fecal microbial flora for anti- 28 STONE HH. Basic principles in the use of prophylactic antibiotics. J Antimicrob biotic resistance in ceftiofur-treated calves. Foodborne Pathog Dis 2006;3:355–65. Chemother 1984;14(Suppl B):33–7. 2 LEITE-MARTINS LR, MAHÚ MI, COSTA AL, et al. Prevalence of antimicro- 29 CLASSEN DC, EVANS RS, PESTOTNIK SL, et al. The timing of prophylactic bial resistance in enteric Escherichia coli from domestic pets and assessment of administration of antibiotics and the risk of surgical-wound infection. N Engl J associated risk markers using a generalized linear mixed model. Prev Vet Med Med 1992;326:281–6. 2014;117:28–39. 30 DUBREUIL P, DAIGNEAULT J, COUTURE Y, et al. Penicillin concentrations in 3 RANTALA M, LAHTI E, KUHALAMPIL J, et al. Antimicrobial resistance in serum, milk, and urine following intramuscular and subcutaneous administra- Staphylococcus spp., Escherichia coli and Enterococcus spp. in dogs given antibi- tion of increasing doses of procaine penicillin G in lactating dairy cows. Can J otics for chronic dermatological disorders, compared with non-treated control Vet Res 2001;65:173–80. dogs. Acta Vet Scand 2004;45:37–45. 31 MEVIUS DJ, NOUWS JF, BREUKINK HJ, et al. Comparative pharmacokinetics, 4 The Review On Antimicrobial Resistance. O'NEILL J, ed. Antimicrobial bioavailability and renal clearance of five parenteral oxytetracycline-20% formu- Resistance: Tackling a crisis for the health and wealth of nations, 2014. http:// lations in dairy cows. Vet Q 1986;8:285–94. www. amr- review. org/. 32 PAPICH MG, KORSRUD GO, BOISON JO, et al. A study of the disposition of 5 Commonwealth Of Australia. National antimicrobial resistance strategy 2015- procaine penicillin G in feedlot steers following intramuscular and subcutaneous 2019. 2016 ht tp:// www.health. gov. au/inter net/ main /pub lish ing. nsf/ Content/ injection. J Vet Pharmacol Ther 1993;16:317–27. 1803C43 3C71 415 CACA257C8400121B1F/$File/ amr- strategy- 2015- 2019. pdf. 33 CHICOINE AL, DOWLING PM, BOISON JO, et al. A survey of antimicrobial 6 United Nations general assembly. Political declaration of the high-level meeting use during bovine abdominal surgery by western Canadian veterinarians. Can of the general assembly on the prevention and control of non-communicable Vet J 2008;49:1105–9. diseases, 2011. (accessed 21 May 2015). 34 YELON JA, GREEN JD, EVANS JT. Efficacy of an intraperitoneal antibiotic to 7 World Health Organisation. Global action plan on antimicrobial resistance. 2015 reduce the incidence of infection in the trauma patient: a prospective, rand- http://www. who. int/ antimicrobial- resistance/ global- action- plan/ en/. omized study. J Am Coll Surg 1996;182:509–14. 8 BOSCH T, VERKADE E, VAN LUIT M, et al. Transmission and persistence of 35 KLEIN WR, FIRTH EC, KIEVITS JM, et al. Intra-abdominal versus intramus- livestock-associated methicillin-resistant Staphylococcus aureus among veterinar- cular application of two ampicillin preparations in cows. J Vet Pharmacol Ther ians and their household members. Appl Environ Microbiol 2015;81:124–9. 1989;12:141–6. 9 DOHMEN W, BONTEN MJ, BOS ME, et al. Carriage of extended-spectrum 36 POLLOCK AV, EVANS M, SMITH GM. Preincisional intraparietal Augmentin -lactamases in pig farmers is associated with occurrence in pigs. Clin Microbiol in abdominal operations. Ann R Coll Surg Engl 1989;71:97–100. Infect 2015;21:917–23. 37 TAYLOR TV, WALKER WS, MASON RC, et al. Preoperative intraparietal 10 LIU W, LIU Z, YAO Z, et al. The prevalence and influencing factors of methicil- (intra-incisional) cefoxitin in abdominal surgery. Br J Surg 1982;69:461–2. lin-resistant Staphylococcus aureus carriage in people in contact with livestock: A 38 NOUWS JF, BREUKINK HJ, BINKHORST GJ, et al. Comparative pharmacoki- systematic review. Am J Infect Control 2015;43:469–75. netics and bioavailability of eight parenteral oxytetracycline-10% formulations 11 Australian Pesticides And Veterinary Medicines Authority. Quantity of in dairy cows. Vet Q 1985;7:306–14. Antimicrobial Products Sold for Veterinary Use in Australia: July 2005 to June 39 PITKÄLÄ A, HAVERI M, PYÖRÄLÄ S, et al. Bovine mastitis in Finland 2010. 2014 http://www. apvma. gov. au/. 2001--prevalence, distribution of bacteria, and antimicrobial resistance. J Dairy 12 DE BRIYNE N, ATKINSON J, POKLUDOVÁ L, et al. Antibiotics used most Sci 2004;87:2433–41. commonly to treat animals in Europe. Vet Rec 2014;175:325. 40 HARADA K, ASAI T, KOJIMA A, et al. Antimicrobial susceptibility of patho- 13 DE BRIYNE N, ATKINSON J, POKLUDOVÁ L, et al. Factors influencing anti- genic Escherichia coli isolated from sick cattle and pigs in Japan. J Vet Med Sci biotic prescribing habits and use of sensitivity testing amongst veterinarians in 2005;67:999–1003. Europe. Vet Rec 2013;173:475. 41 YOSHIMURA H, ISHIMARU M, ENDOH YS, et al. Antimicrobial susceptibility 14 GIBBONS JF, B OLAND F, BUCKLEY JF, et al. Influences on antimicrobial of Pasteurella multocida isolated from cattle and pigs. J Vet Med B Infect Dis Vet prescribing behaviour of veterinary practitioners in cattle practice in Ireland. Public Health 2001;48:555–60. Vet Rec 2013;172:14. 42 ABRAHAM S, JORDAN D, WONG HS, et al. First detection of extended-spec- 15 Australian Strategic And Technical Advisory Group On Antimicrobial trum cephalosporin- and fluoroquinolone-resistant Escherichia coli in Australian Resistance. Importance rating and summary of antibacterials used in human food-producing animals. J Glob Antimicrob Resist 2015;3:273–7. health in Australia: Commonweath of Australia, 2015. http://www.health. gov. 43 BARLOW RS, MCMILLAN KE, DUFFY LL, et al. Prevalence and Antimicrobial au/ internet/ main/ publishing. nsf/ Content/ ohp- amr. htm. Resistance of Salmonella and Escherichia coli from Australian Cattle Populations 16 NELSON RL, GLENNY AM, SONG F. Antimicrobial prophylaxis for colorectal at Slaughter. J Food Prot 2015;78:912–20. surgery. Cochrane Database Syst Rev 2009:CD001181. 44 DE JONG A, THOMAS V, SIMJEE S, et al. Pan-European monitoring of suscep- 17 HARDEFELDT LY, BROWNING GF, T HURSKY K, et al. Antimicrobials used tibility to human-use antimicrobial agents in enteric bacteria isolated from for surgical prophylaxis by companion animal veterinarians in Australia. Vet healthy food-producing animals. J Antimicrob Chemother 2012;67:638–51. Microbiol 2017;203:301–7. 45 MCDONALD M, GRABSCH E, MARSHALL C, et al. Single- versus multi- 18 THINC. Report on Projection Modelling for the Veterinarian Workforce: ple-dose antimicrobial prophylaxis for major surgery: a systematic review. Aust Australian Veterinary Association, 2015. http://www. ava. com. au/. N Z J Surg 1998;68:388–95. 19 BORG H, CARMALT JL. Postoperative septic arthritis after elective equine 46 HARTNACK AK, NIEHAUS AJ, ROUSSEAU M, et al. Indications for and factors arthroscopy without antimicrobial prophylaxis. Vet Surg 2013;42:262–6. relating to outcome after rumenotomy or rumenostomy in cattle: 95 cases 20 HARDEFELDT LY, BROWNING GF, THURSKY K, et al. Antimicrobials used for (1999-2011). J Am Vet Med Assoc 2015;247:659–64. surgical prophylaxis by equine veterinary practitioners in Australia. Equine Vet J 47 KLEIN WR, VAN DER VELDEN MA, ENSINK JM. Single intraoperative 2017. administration of antibiotic to cows with caecal torsion: Wound infection 21 GIBLIN TB, SINKOWITZ-COCHRAN RL, HARRIS PL, et al. Clinicians' percep- and postoperative performance. A retrospective and prospective study. Ve t Q tions of the problem of antimicrobial resistance in health care facilities. Arch 1994;16:111–3. Intern Med 2004;164:1662–8. 48 KLEIN WR, FIRTH EC. Infection rates in clean surgical procedures with and 22 CORTOOS PJ, DE WITTE K, PEETERMANS WE, et al. Opposing expectations without prophylactic antibiotics. Vet Rec 1988;123:542–3. and suboptimal use of a local antibiotic hospital guideline: a qualitative study. J 49 HOLLOWAY S, TROTT DJ, SHIPSTONE M, et al. Antibiotic prescribing Antimicrob Chemother 2008;62:189–95. detailed guidelines, 2013. http://www. ava. com. au/ sites/ default/ files/ AVA_ 23 DE SOUZA V, MACFARLANE A, MURPHY AW, et al. A qualitative study website/ pdfs/ AIDAP guidelines. pdf of factors influencing antimicrobial prescribing by non-consultant hospital 50 British Equine Veterinary Association. PROTECT MEAccess, 2016. doctors. J Antimicrob Chemother 2006;58:840–3. 51 Australian Veterinary Association. Australian Veterinary Workforce Survey. 24 TAN JA, NAIK VN, LINGARD L. Exploring obstacles to proper timing of 2014;2014 http://www. ava. com. au/ workforce- data prophylactic antibiotics for surgical site infections. Qual Saf Health Care 2006;15:32–8. 25 CRIBB A, BARBER N. Prescribers, patients and policy: the limits of technique. Health Care Anal 1997;5:292–8. 26 BURDON DW. Principles of antimicrobial prophylaxis. World J Surg 1982;6:262–7.

6 of 6 | Veterinary Record | 10.1136/vr.104375

83 BACK TO TABLE OF CONTENTS

Chapter 7:

EFFECT OF HISTORIC LABELLING OF ANTIMICROBIALS ON ANTIMICROBIAL STEWARDSHIP IN VETERINARY PRACTICE

8

4 Antimicrobial labelling in Australia: a threat to antimicrobial stewardship? LY Hardefeldta,b*, JR Gilkersona,b, H Billman-Jacobea,b, MA Stevensona,b, K Thurskyc, GF Browninga,b# and KE Baileya,b# aAsia-Pacific Centre for Animal Health, Melbourne Veterinary School, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria, Australia bNational Centre for Antimicrobial Stewardship, Melbourne Veterinary School, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria, Australia cNational Centre for Antimicrobial Stewardship, Peter Doherty Institute, Grattan St, Carlton, Victoria, Australia #These authors contributed equally to this work

This paper has been accepted for publication in the Australian Veterinary Journal.

Abstract Antimicrobial resistance is a public health emergency, placing veterinary antimicrobial use under growing scrutiny. Antimicrobial stewardship, through promotion of the appropriate use of antimicrobials, is a response to this threat. The need for antimicrobial stewardship in Australian veterinary practices has had limited investigation. A survey as undertaken to investigate antimicrobial usage patterns by Australian veterinarians detected the use of varying, and often inappropriate, antimicrobial w dose rates. Doses of procaine penicillin in horses and cattle were often low, with 68% and 90% of respondents reporting doses that were unlikely to result in plasma concentrations above minimum inhibitory concentrations for common equine and bovine pathogens. Frequency of penicillin administration was also often inappropriate. Gentamicin doses in horses were largely appropriate (89% of dose rates appropriate), but 9% of respondents reported twice daily dosing. Amoxycillin and amoxycillin / clavulanate were administered at the appropriate doses, or above, to dogs and cats by 54% and 70% of respondents respectively. In this commentary, we have explored the potential reasons for inappropriate antimicrobial dose regimes and have found that antimicrobial labels often recommend incorrect dose rates, and thus may be contributing to poor prescribing practices. Mechanisms to ensure that antimicrobial drug labels are regularly updated to reflect the dose needed to effectively and safely treat common veterinary pathogens should be adopted. This will especially be true if changes in legislation restrict antimicrobial use by veterinarians to the uses and doses specified on the label. This change would hamper the current momentum towards improved antimicrobial stewardship.

Abbreviations AMR; antimicrobial resistance

Introduction Antimicrobial resistance (AMR) is regarded to be a public health emergency.1 In 2016, a review on AMR commissioned by the United Kingdom government suggested that approximately 700,000 people have died because of multi-drug resistant infections, and predicted that by 2050 the mortality rate may exceed 10 million people.2 Antimicrobials are an essential part of veterinary medicine, and their use ensures both the welfare of animals under veterinary care and the ongoing security and safety of food. However, the misuse of antimicrobials cannot be justified. Antimicrobial stewardship encompasses the full range of measures used to promote appropriate antimicrobial prescribing and reduce the selective pressure for development of AMR.

8

5 In 2016, a survey was undertaken to investigate antimicrobial usage patterns by Australian veterinarians.3-5 In total, 721 members of the veterinary profession completed the companion animal practice questions, 337 completed the equine practice questions and 212 completed the bovine practice questions. The questionnaire respondents used predominantly low-importance rating antimicrobials, as defined by the Australian Strategic Technical Advisory Group on AMR.6 The classes of antimicrobials most frequently administered for prophylaxis during small animal surgery were aminopenicillins (53%), predominantly amoxicillin / clavulanate (40%), and first generation cephalosporins (36%).3 For equine surgery, the most common class of antimicrobials administered was penicillins, with procaine penicillin selected by 66% of respondents and long-acting penicillins selected by 3.3% of respondents.4 Other antimicrobials commonly administered to horses were trimethoprim-sulphonamide (17%) and gentamicin (12%).4 In cattle, penicillins (39% procaine penicillin, 5% long- acting penicillin formulations) and tetracyclines (46% oxytetracycline) were the antimicrobials most commonly administered for surgical prophylaxis.5

There are multiple sources of information about dose rates that may be used by practitioners, including educational and training materials, authoritative text books and the label recommendations provided by the manufacturer of the drug (and approved by the Australian Pesticides and Veterinary Medicines Authority).3-5 At the time of the survey, guidelines for antimicrobial use were available for companion animals,7, 8 but not for horses or cattle, although guidelines for these species have been released recently in Australia.8 The current best-practice dose rate for procaine penicillin required to achieve appropriate plasma concentrations of penicillin in horses is 22,000 IU/kg every 12 hours,9-11 and in cattle 22,000 IU/kg every 12 to 24 hours.12 However, the label for procaine penicillin for both cattle and horses recommends an intramuscular dose of 20 ml/500 kg (12,000 IU/kg) every 24 hours.13 Similarly, current recommended dose rates for gentamicin in horses are 7.7-9.7 mg/kg once per day,14 but the labels for gentamicin recommends a dose of 1.5mg/kg15 or 2 mg/kg16 three times per day. The current best evidence suggests that dose regimens for amoxycillin of 15 mg/kg every 12 hours17 and for amoxycillin/clavulanate of 12.5 mg/kg every 12 hours18 are appropriate for dogs, but the labels recommend 7 mg/kg19 and 8.75 mg/kg,20 respectively.

This commentary aims to explore antimicrobial dose rates used by Australian veterinarians for small animals, cattle and horses based on surveys on antimicrobial usage patterns by Australian veterinarians.

The survey method has been described previously.3 Surveys were undertaken using an online questionnaire to collect information about the respondents, the frequency with which antibiotics were used for specific surgical conditions (including the dose, timing and duration of therapy) and practice antimicrobial use policies and sources of information about antimicrobials and their uses.3 An aspect of the results that has not been reported previously was the range of dose rates used by equine and companion animal veterinarians.

Equine practitioners prescribed procaine penicillin in doses that varied widely, with 12,000, 15,000 and 22,000 IU/kg most frequently reported (Figure 1). Doses reported

8

6 were less than the recommended dose in 65% of responses. Procaine penicillin was administered twice daily by 74% of respondents and once daily by 26% of respondents. Trimethoprim/sulphonamide was administered most commonly at 15 or 30 mg/kg (Figure 1) and was administered twice daily by 94% of respondents, with 6% reporting once daily dosing. Reported trimethoprim/sulphonamide dose rates were less than the recommended dose in 70% of scenarios. Gentamicin was most commonly administered at 6.6 mg/kg (Figure 1) and was administered once daily by 92% of respondents, with 9% reporting twice daily dosing. Only 6% of respondents indicated using gentamicin at a dose lower than that commonly recommended.

Companion animal veterinarians most commonly prescribed amoxycillin at a dose rate of 15 mg/kg (Figure 2) and the frequency of administration was most commonly twice daily (99%). Doses reported were lower than the recommended dose in 46% of scenarios. Amoxycillin/clavulanate was most commonly administered at 12.5 mg/kg and 8.75 mg/kg (Figure 2) and the frequency of administration was most commonly twice daily (99%). Reported doses were lower than the recommended dose in 29% of scenarios.

Dose rates used by bovine practitioners have previously been reported.5 The dose of procaine penicillin administered ranged from 7.5 - 24 mg/kg, with the most frequent dose rates being 12.5 mg/kg and 15 mg/kg.5 Reported doses were lower than the recommended dose in 90% of scenarios. Procaine penicillin was administered once daily by 96% of respondents using a short acting formulation, and twice daily by 5% of respondents.5 Similarly, for oxytetracycline the dose range was wide (2.5 - 13.5 mg/kg), with the most frequent dose rates being 4-5 mg/kg and 10 mg/kg.5 Reported doses were lower than the recommended dose in 59% of scenarios. Most respondents reported administering oxytetracycline once daily when using a short acting

8

7 formulation (97%), with the remainder reporting that they administered it every 12 hours (2.8%).5

Discussion The survey identified a wide range of dose rates in use for common antimicrobials used in horses, consistent with previous findings described in cattle.21 Of concern is a predominant use of dose rates lower than those required to result in plasma concentrations of the antimicrobial above the minimum inhibitory concentrations required for efficacy against common equine and bovine pathogens.22 The recommended dose and inter-dosing interval for horses and cattle are shown in Tables 1 and 2, respectively. Comparatively, the dose rates in use for common antimicrobials used in companion animal medicine were mostly at or above those recommended in the literature.20, 23

Poor teaching of veterinary undergraduates is one potential reason for failure of veterinarians to administer appropriate doses of antimicrobial agents. Veterinary pharmacology is taught as a stand-alone subject in 5 of the 7 veterinary schools in Australia24, 25 (personal communication CS, MW, SR, JM, CP, RC, TN, 2017). It is integrated into other courses at the University of Melbourne and James Cook University

8

8 (personal communication RT, 2017). Integration of pharmacology may lead to reduced emphasis and requirement for knowledge in this area. However, appropriate drug doses are taught at all Australian veterinary schools,4 which is consistent with the data showing Australian university equine veterinarians had 3.2 times higher odds of compliance with guidelines than general practitioners in the 2016 survey.4 Thus, it seems unlikely that inappropriate university teaching is the primary reason for under- dosing of antimicrobial agents in equine and bovine practice.

A second potential contributor to inappropriate dosing is the labelling of antimicrobials. Some of the doses recommended on the labels for antimicrobials fall below those now recognised as appropriate. Recommended doses on the labels for procaine penicillin, gentamicin, amoxycillin and amoxycillin/clavulanate are all lower than those currently recommended in the literature.9-11, 14, 17, 18 In addition, the label for gentamicin in horses recommends administration three times per day, which is likely to increase the risk of renal toxicity in treated horses.26 The gentamicin label also carries a “do not use” clause that prohibits use unless sensitivity testing indicates that other antimicrobials are ineffective. In most jurisdictions in Australia, this clause is legally binding and off-label use in the absence of sensitivity data is, therefore, illegal. Despite this, the majority of practitioners appear to use gentamicin as part of a broader spectrum empirical approach to therapy, including for prophylaxis in compromised surgical scenarios. While this use would clearly be contrary to the off-label use restrictions, as culture and sensitivity have not been performed prior to use, it is a preferable approach to administration of third generation cephalosporins or fluoroquinolones as these are both antimicrobial classes with a high importance rating. Thus, this label restriction could drive practitioners towards less appropriate therapeutic options if enforced. Inappropriate labels are likely to influence bovine prescribing where the withholding periods for meat and milk declared on the label are only applicable for the labelled dose. Improvements in accuracy of labels would enable the use of appropriate doses in this species.

While the dose recommended by the label is likely to be a factor influencing some veterinarians, the label is clearly not the only factor leading to inappropriate antimicrobial dosing. Respondents to our equine survey frequently used off-label (appropriate) doses of gentamicin in horses (89% of dose rates appropriate), but often used labelled (inappropriate) doses of procaine penicillin (32% of dose rates appropriate). The doses of amoxycillin/clavulanate also reflect this, with 23% of companion animal practitioners using the labelled (inappropriate) dose and 50% using an off-label (appropriate) dose. A change in legislation to make off-label use of antimicrobials illegal, without ensuring regular updates to the labelling of veterinary medicines, would clearly be detrimental to the aim of improved antimicrobial stewardship.

The risk of side-effects associated with using gentamicin at the labelled dose and inter- dosing interval may have led to an emphasis on the appropriate dosing of this drug, whereas the risk of under-dosing procaine penicillin and oxytetracycline is insidious. However, the use of amoxycillin and amoxycillin/clavulanate is largely at an appropriate (off-label) dose and risks of under-dosing of these drugs are also not immediately obvious. It is possible that these differences between companion animal and large animal practitioners are influenced by differences in the source of continuing

8

9 education or differing opportunities to attend continuing education. The culture within veterinary practices may also contribute to the dose rates used by veterinarians. This may be especially true for recent graduates, who are likely to rely more heavily on the experience of colleagues to guide clinical decision making, or for veterinarians who are not predominantly working in equine or bovine practice and for whom dose rates for these species may not be easily recalled. However, we have not yet evaluated the role of these factors.

Conclusion Mechanisms to ensure that antimicrobial drug labels are regularly updated to reflect the dose needed to effectively and safely treat common veterinary pathogens should be adopted, which would also ensure that label restrictions are not driving selection of inappropriate antimicrobials for therapy. This will especially be true if changes in legislation restrict antimicrobial use by veterinarians to that directed by the label. This change would hamper the current momentum towards improved antimicrobial stewardship. Development of evidence-based antimicrobial use guidelines may also improve dose rates used by veterinarians in Australia.

References 1. World Health Organisation. United Nations high-level meeting on antimicrobial resistance. 2016. Retrieved 13/9/17 http://www.who.int/antimicrobial- resistance/events/UNGA-meeting-amr-sept2016/en/. 2. O'Neill J. Tackling drug-resistant infections globally: Final report and recommendations. The Review on Antimicrobial Resistance, https://amr- review.org/sites/default/files/160525_Final paper_with cover.pdf, 2016. 3. Hardefeldt LY, Browning GF, Thursky K et al. Antimicrobials used for surgical prophylaxis by companion animal veterinarians in Australia. Veterinary Microbiology 2017;203:301- 307. 4. Hardefeldt LY, Browning GF, Thursky K et al. Antimicrobials used for surgical prophylaxis by equine veterinary practitioners in Australia. Equine Vet J 2017. 5. Hardefeldt LY, Browning GF, Thursky KA et al. Cross-sectional study of antimicrobials used for surgical prophylaxis by bovine veterinary practitioners in Australia. Vet Rec 2017. 6. Australian Strategic and Technical Advisory Group on Antimicrobial Resistance. Importance rating and summary of antibacterials used in human health in Australia. Commonweath of Australia, http://www.health.gov.au/internet/main/publishing.nsf/Content/ohp-amr.htm, 2015. 7. Holloway S, Trott DJ, Shipstone M et al. Antibiotic Prescribing: detailed guidelines. Australasian Infectious Diseases Advisory Panel, www.ava.com.au/sites/default/files/AVA_website/pdfs/AIDAP guidelines.pdf, 2013. 8. Asia Pacific Centre for Animal Health, National Centre for Antimicrobial Stewardship. Australian Veterinary Prescribing Guidelines. www.fvas.unimelb.edu.au/vetantibiotics. 2017. Retrieved 13/9/17. 9. Uboh CE, Soma LR, Luo Y et al. Pharmacokinetics of penicillin G procaine versus penicillin G potassium and procaine hydrochloride in horses. Am J Vet Res 2000;61:811-815. 10. Firth EC, Nouws JF, Driessens F et al. Effect of the injection site on the pharmacokinetics of procaine penicillin G in horses. Am J Vet Res 1986;47:2380-2384. 11. Love DN, Rose RJ, Martin IC, Bailey M. Serum concentrations of penicillin in the horse after administration of a variety of penicillin preparations. Equine Vet J 1983;15:43-48. 12. Dubreuil P, Daigneault J, Couture Y, Guay P, Landry D. Penicillin concentrations in serum, milk, and urine following intramuscular and subcutaneous administration of increasing doses of procaine penicillin G in lactating dairy cows. Can J Vet Res 2001;65:173-180.

90 13. MSD Animal Health. Depocillin Procaine Penicillin 300ng/ml Injection. www.infopest.com.au. Retrieved 10/3/17. 14. Bauquier JR, Boston RC, Sweeney RW, Wilkins PA, Nolen-Walston RD. Plasma peak and trough gentamicin concentrations in hospitalized horses receiving intravenously administered gentamicin. J Vet Intern Med 2015;29:1660-1666. 15. Nature Vet PtyLtd. Gentamax 100. http://websvr.infopest.com.au/LabelRouter?LabelType=L&Mode=1&ProductCode=5587 3. Retrieved 15/9/17. 16. Troy Laboratories PTY LTD. Gentam 100. www.infopest.com.au. Retrieved 26/4/17. 17. ten Voorde G, Broeze J, Hartman EG, van Gogh H. The influence of the injection site on the bioavailability of ampicillin and amoxycillin in beagles. Vet Q 1990;12:73-79. 18. Bywater RJ, Palmer GH, Buswell JF, Stanton A. Clavulanate-potentiated amoxycillin: activity in vitro and bioavailability in the dog. Vet Rec 1985;116:33-36. 19. Jurox. Moxylan ready-to-use injection 150mg/ml. www.infopest.com.au. Retrieved 1/6/17. 20. Norbrook Laboratories Australia. Noroclav injection for dogs and cats. Retrieved 1/6/17. 21. Hardefeldt LY, Browning GF, Thursky K et al. Antimicrobials used for surgical prophylaxis by bovine veterinary practitioners in Australia. Veterinary Record 2017;In press. 22. Prescott JF. Beta-lactam antibiotics: Penam Penicillins. In: Giguere S, Prescott JF, Dowling PM, editors. Antimicrobial Therapy in Veterinary Medicine. 5th edn. Wiley Blackwell, Iowa, USA, 2013. 23. Norbrook Laboratories Australia. Betamox injection. www.infopest.com.au. Retrieved 1/6/17. 24. University of Queensland. Bachelor of Veterinary Science Study Planner. 2017. Retrieved 22/9/17. 25. University of Adelaide. Doctor of Veterinary Medicine. 2017. Retrieved 22/9/17. 26. van der Harst MR, Bull S, Laffont CM, Klein WR. Gentamicin nephrotoxicity--a comparison of in vitro findings with in vivo experiments in equines. Vet Res Commun 2005;29:247- 261. 27. Gustafsson A, Baverud V, Franklin A et al. Repeated administration of trimethoprim/sulfadiazine in the horse--pharmacokinetics, plasma protein binding and influence on the intestinal microflora. J Vet Pharmacol Ther 1999;22:20-26. 28. Mevius DJ, Nouws JF, Breukink HJ et al. Comparative pharmacokinetics, bioavailability and renal clearance of five parenteral oxytetracycline-20% formulations in dairy cows. Vet Q 1986;8:285-294.

91 BACK TO TABLE OF CONTENTS

Chapter 8:

ENABLERS OF, AND BARRIERS TO, ANTIMICROBIAL STEWARDSHIP IN VETERINARY PRACTICE

9

2 The barriers to, and enablers of, implementing antimicrobial stewardship programs in veterinary practices LY Hardefeldta,b, JR Gilkerson, H Billman-Jacobea,b, MA Stevensona, K Thurskyb, KE Baileya,b and GF Browninga,b a Asia-Pacific Centre for Animal Health, Melbourne Veterinary School, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria, Australia b National Centre for Antimicrobial Stewardship, Peter Doherty Institute, Grattan St, Carlton, Victoria, Australia This paper is under consideration for publication.

Abstract Background: Antimicrobial stewardship programs are yet to be widely implemented in veterinary practice and medical programs are unlikely to be directly applicable to veterinary settings. Objective: To gain an in-depth understanding of the factors that influence effective antimicrobial stewardship in veterinary practices in Australia. Methods: A concurrent explanatory mixed methods design was used. The quantitative phase of the study comprised an online questionnaire to assess veterinarians’ attitudes to antimicrobial resistance and antimicrobial use in animals, and the extent to which antimicrobial stewardship is currently implemented (knowingly or unknowingly). The qualitative phase used semi-structured interviews to gain an understanding of the barriers to and enablers of antimicrobial stewardship in veterinary practices. Data were collected and entered into NVivo v.11, openly coded and analysed according to mixed methods data analysis principles. Results: Companion animal, equine and bovine veterinarians participated in the study. Veterinary practices rarely had antimicrobial prescribing policies. The key barriers were a lack of antimicrobial stewardship governance structures, client expectations and competition between practices, the cost of microbiological testing, and a lack of access to education, training and antimicrobial stewardship resources. The enablers were concern for the role of veterinary antimicrobial use in development of antimicrobial resistance in humans, a sense of pride in the service provided, firstly, and preparedness to change prescribing practices. Conclusion and clinical importance:secondly, This st udy can guide the development and establishment thirdly, of antimicrobial stewardship programs in veterinary practices, by defining the major issues that influence the prescribing behaviour of veterinarians.

Introduction Antimicrobial resistance (AMR) is a global health emergency. Use of antimicrobials in animals has been implicated in the emergence of AMR in bacterial populations, with undesirable consequences for both human and animal health1,2. Antimicrobial stewardship (AMS) programs are widely implemented in human hospitals worldwide and have been shown to improve clinical outcomes for patients while limiting the emergence and spread of AMR3. Global4 and national5 strategies for tackling AMR have called for improved AMS in veterinary practices, but there have been no formal reports describing the outcomes of AMS programs that have been implemented in veterinary practices to date.

9

3 Medical strategies for AMS are unlikely to be directly applicable to the veterinary context, in part because of the differences in the availability of human and financial resources for the diagnosis and treatment of individual animals, geographical spread and limited tools supporting AMS in the veterinary sector. Most veterinary practices in Australia employ fewer than 5 veterinarians (87% in 2000) and the average profit margin is 16%6. Importantly, this profit is inclusive of profit from dispensing of pharmaceutical agents. The veterinary profession will need to develop strategies for AMS that are innovative and appropriate to the size, variability, and resource availability of the majority of veterinary practices. The aim of this study was to gain an in-depth understanding of the factors that influence AMS in Australian veterinary practices.

Materials and methods This was a cross-sectional study to assess veterinarians’ attitudes to antimicrobial resistance and antimicrobial use in animals in Australia. A concurrent explanatory mixed methods design was used, in which a preliminary quantitative process contributed to a principally qualitative study7,8. The quantitative phase comprised an online questionnaire to assess veterinarians’ attitudes to AMR and antimicrobial use in animals, and the extent to which AMS is currently implemented (knowingly or unknowingly) in their practice. The qualitative phase consisted of semi-structured interviews to understand the barriers to and enablers of AMS in veterinary practices in Australia. This design allowed a study of specific aspects of AMS, with exploration of the original themes, in a range of veterinary practice types, with triangulation of the findings to ensure consistency.

Quantitative An on-line questionnaire was developed with both open and closed questions (questionnaire available as supplementary information) asking veterinarians to provide details of their attitudes to antimicrobial resistance and antimicrobial use in animals and the extent to which antimicrobial stewardship is currently implemented (knowingly or unknowingly) in their area of practice. The questionnaire was sent to practices participating in the qualitative survey. Announcements were made using social media, and responses were requested at the Australian Veterinary Association Conference (Melbourne, June 2017) between February and June 2017.

Sample size calculations were performed to determine the number of respondents required to make appropriate inferences from the survey. To be 95% certain that our estimate of the population prevalence of veterinarians using a given class of antimicrobial was within 7.5% of the true population prevalence, a total of 168 completed surveys were required (10,000 veterinarians were estimated to be practicing in Australia at the time of the survey). The entire questionnaire took about 10 minutes to complete, encompassed four question areas with a maximum of 36 questions in total. The questionnaire was trialled with four general practitioners unaffiliated with the research team, and modified iteratively to improve clarity, face validity and content validity. Descriptive statistics were used to summarise participants’ data.

9

4 Qualitative A qualitative approach involving semi-structured interviews with veterinarians was employed. Interview themes were developed using the COM-B framework9. A purposive sample approach was used to select participants to ensure inclusion of a diverse range of clinical practice (Figure 1). Participants were recruited until there was a diverse range of practice type and data saturation was reached on thematic analysis.

The semi-structured interview guide was informed by a literature review and findings from previous surveys and was piloted with two veterinarians. There were three key areas addressed: attitudes to and experiences of AMR, current AMS processes and needs for and barriers to proposed components of AMS programs. Between March and June 2017 face-to-face interviews were conducted at the veterinary clinics involved. Informed consent was provided and the interviews were audio recorded with participants’ consent. The interviews lasted, on average, 45-60 minutes and were conducted by one author (LYH).

Interviews were audiotaped, transcribed, entered into NVivo version 11 (QSR International), and openly coded and analysed by one researcher (LYH) using qualitative data analysis principles10-13 and thematic analysis14. A second researcher (GFB) independently analysed two of the transcripts to ensure reliability. Identified inconsistencies in the codes were discussed and the themes generated were agreed upon. The code structure was developed using an inductive approach. Special attention was paid to any notable variation between veterinarians from metropolitan and rural areas, and those in companion animal-only practice versus those in equine or cattle practice (with or without a companion animal component), and between practice owners or directors and employees.

Ethical clearance This research was approved by the University of Melbourne Faculty of Veterinary and Agricultural Sciences Human Ethics Advisory Group under Approval No. 1648135.1.

Results A total of 184 responses to the on-line questionnaire were received. The demographics of the respondents are presented in Table 1. Veterinary practices were recruited for focus group interviews until data saturation was reached on thematic analysis, which occurred after interview seven, and continued to ensure consistency across a diverse range of practice. A total of 13 interviews were conducted (Figure 1). Information on participants recruited for the interviews are presented in Table 1 and were similar to the survey respondents and the national workforce, where data exists. A coding tree was designed (Supplementary table 1) and based on emerging themes the enablers of and barriers to AMS in veterinary practices were explored. We found no discernible difference in experiences between rural and metropolitan practices, nor between practice owners and employees. Differences were found between veterinarians working in companion animal only practices and those in practices serving a clientele that owned horses or cattle. The key findings are summarized in Table 2.

9

5 Figure 1. Qualitative study logistics.

Table 1. Demographics of survey respondents and interview participants compared to national veterinary workforce Australian veterinary Survey Interview workforce respondents participants 6,15 Characteristic N (%) N (%) % Gender Male 62 (36) 10 (26) 39 Female 111 (64) 29 (74) 61

Location Capital city 76 (42) 19 (49) 50 Other 105 (58) 20 (51) 50

Years in practice 0-5 40 (23) 11 (28) NA 6-15 63 (36) 19 (49) NA >15 73 (41) 9 (23) NA

Position in practice NA Owner/director 11 (28) NA Associate 28 (72) NA

Type of practice Companion animal 92 (51) 20 (51) NA only 87 (49) 19 (49) NA Equine or Bovine +/- companion animal NA, not available

Key themes identified Perceptions of AMR Multi-drug resistant pathogens (MDR) were rarely encountered by survey respondents (88% [162/184] reported encountering MDR pathogens less frequently than monthly or never). The most commonly encountered MDR pathogens were extended spectrum beta-lactamase producing gram negative organisms, methicillin-resistant

9

6 Staphylococcus aureus and methicillin-resistant Staphylococcus pseudintermedius (54%, 32% and 21% of the 159 respondents who reported culturing MDR pathogens, respectively). Similarly, antimicrobial resistance was infrequently encountered by most focus group participants, with only some reporting frequently dealing with multi-drug resistant infections. Many reported they

Table 2. Summary of major barriers and enablers for implementing AMS programs in veterinary practices Major barriers Major enablers Client expectations & competition between Concern for human health practices Cost of microbiological testing Pride in service provided Lack of access to education & training Low level of resistance encountered Lack of AMS governance structures Preparedness to change prescribing practices Lack of independent guidelines for antimicrobial Frequent use of low cost diagnostic tests use Hierarchical structure of many practices Low use of most critically important antimicrobial agents felt that the medical profession, and in particular medical general practitioners, were most responsible for AMR in humans in Australia. However, veterinarians frequently reported feeling partly responsible for AMR in human medicine and were often concerned about that contribution. For example: “I'm worried about my influence, if I'm causing it. And then I guess I do know people that have had elective procedures that have developed resistance. So I don't want to add to it.”

In addition, there were a wide range of opinions about the significance of the effect of antimicrobial use in different species on the risk of AMR in humans. Over 50% (94/184) of respondents to the questionnaire indicated that veterinary antimicrobial use had a moderate contribution to overall AMR, but over 60% (111/184) indicated that their own antimicrobial use made only a minimal contribution to AMR (Figure 2). The focus group interviews gave greater depth of understanding to this issue. Although most companion animal veterinarians admitted to the overuse of antimicrobials, they thought that antimicrobial use in the dairy and intensive animal industries was most to blame, whereas veterinarians treating dairy cattle attributed most of the risk to the intensive animal industries. For example: “I think statistically it's got to be the human doctors. And we are only a small percentage of prescribers, particularly small animals. And then there's also, there's the growing concern about the production animals and the use of that, but I think that's coming well into focus now as well.”

“Because I think the chance that small animal drugs getting into the human food chain, or antibiotic resistance chain, are much less than food production.”

9

7 70

60

50

40

30

% respondents 20

10

0 None Minimal Moderate Strong Unsure

Contribution of individual veterinarian's antimicrobial use to AMR in humans

Contribution of overall veterinary antimicrobial use to AMR in humans

Figure 2. Proportions of survey respondents indicating how much antimicrobial use by individuals, and by the profession, contributes to the overall burden of AMR.

Some veterinarians treating horses and cattle felt that antimicrobial use in those species was contributing to AMR to some degree. For example: “We do prescribe a lot of antimicrobials, particularly to dairy cows. So that's got to contribute somewhere, I would have thought.”

Some participants felt that there were limited detrimental effects of antimicrobial use in animals. For example: “Lots of vets don’t have any fear of antibiotics, they kind of figure, "Well, if I'm going to give the patient something, then antibiotics would be it."”

Fear of AMR affecting the ability to effectively treat clinical veterinary cases in the future was expressed, as was the potential for dispensing rights of veterinarians being removed as a result of perceived irresponsible use by veterinarians.

Use of critically important antimicrobials16 varied between classes of drugs (Figure 3). Use of 3rd generation cephalosporins was most common (88/184, 48% of respondents indicating at least weekly use), whereas use of fluoroquinolones (43/183, 23%) and other critically important antimicrobials was less common (31/183, 17% of respondents indicating at least weekly use). Most veterinarians who responded to the questionnaire strongly disagreed that profit made from the sale of antimicrobials influenced their decision to prescribe (72% strongly disagreed, 23% disagreed, 4% neither agreed nor disagreed and 1% agreed, of 172 respondents). Consistent with this, only 35% (64/184) of respondents were aware of the amounts of antimicrobials sold by the practice.

9

8 45 40 35 30 25 20

% respondents 15 10 5 0 Daily Twice Weekly Monthly Less Never weekly frequently

3rd generation cephalosporins Fluoroquinolones Other HIRA

Figure 3. Frequency of use of antimicrobials with a high-importance rating HIRA; high-importance rating antimicrobials.

Client expectations of antimicrobial therapy There was uniformity among participants that clients presenting animals to veterinarians expected some form of treatment from them, often antimicrobials. A subset of clients demanded antimicrobials, sometimes without a formal consultation. Most reported that these expectations have a subsequent impact on their prescribing practice, although some felt that it did not influence their therapeutic choices. Antimicrobial dispensing to bona fide clients without formal consultation was not reported in companion animal practice, but was commonly reported in horse and cattle practice. It was deduced that antimicrobials were given without a formal consultation for three reasons. Firstly, veterinarians feel pressured to ‘keep clients happy’ due to competition between practices and the fear that clients would consult a new practice if they were not pleased with the service they received. This was equally common in equine and cattle practice, where clients often have many animals and can contribute proportionately more to the practice’s profitability than an individual companion animal client with only one animal. For example: “Because you have to provide a business and you have to keep the clients happy. That's that difference between us and the medical profession. They still get paid at the end of the day. But if we don't have clients, we don't get paid or have a job. So at some point you do have to keep them happy.”

The second reason was that some clients felt they are capable of diagnosing common diseases and were not willing to pay for a veterinary consultation for routine disease management, and that veterinarians felt that they were unable to examine every animal requiring antimicrobial therapy. However, participants often conceded that the therapy they advise was different from the client’s first preference and that, in many cases, the antimicrobials were not used in accordance with advice given, or with the label, and that consultation with clients was usually led to more appropriate therapy. For example: “But obviously, you're not going out to see every case”

9

9 Finally, veterinarians often reported that, due to long work days and lack of time, it was easier to dispense antimicrobials than to spend time convincing clients that the antimicrobials were not necessary or that a veterinary consultation was required. For example: “At the end of a long day, it's hard to deal with that stuff”

Costs associated with diagnostic testing The factors that influenced the decision to perform culture and susceptibility testing were consistent among the questionnaire respondents; persistent or recurrent infections were the most common reason (74% of respondents [137/184]), while cost constraints of the client (34% of respondents [62/184]), the location of the disease (30% of respondents [56/184]), severe infections (27% of respondents [50/184]), and atypical findings on in-house cytology (18% of respondents [33/184]) were also reported. Diagnostic tests of low cost, such as cytology, were unanimously used by focus group participants in companion animal practice, but less so in horse and cattle practice. Most focus group participants reported that the costs of diagnostic testing, and particularly culture and sensitivity testing, led to an overuse of antimicrobials in their practice. This occurred when a treatment trial with antimicrobials replaced the use of diagnostic tests to investigate the presence of an infection, or when the cost constraints of the client prevented diagnostic testing but the veterinarian feared the consequences that a failure to treat an unlikely infection may have for the health of the animal. For example: “because of the unwillingness of people to necessarily take diagnostic steps. And so it's often offered as a, "Well, we can trial antibiotics and see if it gets better."”

“but there is this fear of what if I neglect to treat something that I should have treated?”

Others felt that they were not under-utilizing diagnostic testing due to the costs of these tests. For example: “I think if it's necessary I'd do it, but I just don't think it's necessary most times.”

Lack of resources Antimicrobial prescribing policies and antimicrobial stewardship policies were uncommon, with only 15% (27/184) of survey respondents indicating that their practice had either of these documents (70% did not have either document, 15% were unsure). For respondents that had access to antimicrobial prescribing policies, 44% (10/23) commented that the policy documented had been created in the past year. None of the focus group participants were practicing in a clinic that had a formal antimicrobial use policy. Guidelines were used by only 28% (51/184) of respondents to the questionnaire, with the most commonly used being the Australasian Infectious Disease Advisory Panel guidelines17 (45% [23/51]) and the British Small Animal Veterinary Association guidelines18 (20% [10/51]). Many focus group participants had access to guidelines, but skepticism was expressed about the involvement of a pharmaceutical company in the production of the antimicrobial use guidelines currently available for companion animals17. For example: “I sort of think that it’s tainted information. Good information that you have to try and sort of filter a little bit. It would be really nice to have a guideline that wasn't sponsored by someone who had something to earn.”

100 Specialist veterinarians, in either internal medicine or surgery, were primarily consulted for advice on clinical cases when colleagues or employers were either not available or were unsure, although rarely for advice on which antimicrobial was most appropriate. However, frequently participants reported relying on personal experience when deciding on antimicrobial therapy and recently graduated veterinarians reported relying on the experience of colleagues. For example: “I still use a little bit of what I thought I knew but it's more attractive to use what everyone else here has done and their protocols”

Antimicrobial stewardship policies were strongly supported by the respondents to the survey, with 89% (163/184) reporting that they felt their practice should have an AMS policy to improve responsible prescribing (40% [65/163]), reduce AMR in animals (16% [26/163]), reduce AMR in humans (15% [26/163]) or because it represents best- practice (10% [17/163]). Common reasons for not having AMS policies in veterinary practice were a misunderstanding of AMS (25% [5/20]), being in solo practice (25% [5/20]) or because they already had low rates of antimicrobial prescribing (25% [5/20]). The most commonly selected factors limiting AMS in practice were pressure from clients, practice culture, client finances and lack of continuing veterinary education (24%, 19%, 19% and 11% of 97 respondents, respectively). Exposure to some form of education about antimicrobial stewardship or appropriate antimicrobial prescribing was reported by 45% (82/184) of respondents, with Australian Veterinary Association national conferences or division meetings reported to be the most frequent source of education (27% [22/82]), followed by self-directed education (23% [19/82]) and webinars or podcasts (16% [13/82]). Additional education was strongly supported by survey respondents (176/183, 96%) and willingness to change prescribing habits based on additional education was also frequently indicated (169/175, 97%).

Discussion To the authors’ knowledge, this is the first study of the enablers of and barriers to AMS in veterinary practice although there have been several studies that have examined attitudes and knowledge about AMR and the impact of antibiotic use19-21. Our results show that 89% of the veterinarian that responded to the questionnaire self-reported that they would support AMS programs in their practices and that limiting factors commonly involve pressure from clients to dispense antimicrobials. This is in contrast to a survey of factors influencing prescribing in European veterinarians, where owner demands were among the least important factors21. However, the interviews indicated that pressure from clients is just one of the factors driving prescribing, and that the situation is complex, with a multitude of contributing influences reflecting the competitive nature of veterinary practice and underlying client-related socioeconomic and situational factors. Instituting AMS programs in veterinary practices is critical, but regulation requiring such programs may be required to overcome the most important barriers of commercial competition between practices and pressure from clients to dispense antimicrobials, often without formal consultation. AMS programs have been widely implemented in human medicine, with key goals of improving and sustaining appropriate antimicrobial prescribing22-25. However, AMS program implementation is not only a challenge facing veterinarians in Australia. Additional challenges remain in promoting sustainable antimicrobial use in human medicine, requiring behavioral change interventions26. In addition, there is evidence from human medicine that

101 interventions that focus on behavioral change can improve antimicrobial prescribing3,26- 28.

Behaviour can be understood to result from an interaction between capability, opportunity and motivation9 and, although the framework has not been specifically assessed for its appropriateness for AMS interventions, it forms a useful platform to interrogate the enablers and barriers in this population of veterinarians. Capability is the physical and psychological skill to institute AMS programs. While awareness of AMS as a movement was widespread, there were still some veterinarians who were unsure about what AMS was, which in itself is a psychological barrier. In addition, lack of education that would enable AMS and costs associated with culture and susceptibility testing were frequently identified as barriers by participants in both facets of this project. These factors represent barriers to AMS capability.

The second part of the framework is opportunity, which encompasses the physical resources and social support needed to institute AMS programs. Formal AMS programs have yet to be instituted in Australian veterinary practices and many of the constituents of these programs are yet to be developed in Australia. At the time the questionnaire was administered, guidelines for antimicrobial use were only available for companion animal practice, and skepticism about the reliability of these guidelines was commonly mentioned in the interviews because of the involvement of a pharmaceutical company in their production. In addition, there is currently no education campaign specifically targeting AMS or appropriate antimicrobial use in Australia. These factors all represent physical barriers to the opportunity for behavioral change. The high levels of interest and support for AMS seen in this project suggest that there is significant social opportunity for AMS.

Motivation is the final part of the framework for behavioral change. Motivation can be reflective, based on one’s conception of self or higher priorities, or automatic, involving emotions and impulses that arise from associative learning or innate dispositions. Participants in both parts of this project exhibited reflective motivation in favor of AMS, as has been found in veterinarians in the United Kingdom29. Few participants reported frequently culturing MDR pathogens, but most felt that the profession had a responsibility to address inappropriate antimicrobial use. Most veterinarians were cognizant of the potential role that veterinary antimicrobial prescribing could play in the development of antimicrobial resistance and most also admitted that overuse of antimicrobials was common in veterinary medicine. Pressure from clients, the fear of negative commercial outcomes, and perceptions that individual contributions to AMR were low adversely affected motivation, as has also been found in general medical practitioners in Australia30.

There are several features of this study that may have influenced the results. Enrolment bias may occur with such surveys as respondents are self-selected. However, respondent demographics were broadly representative of the Australian veterinary profession. In addition, recruitment for the interviews was predominately from practices that expressed an interest in AMS. This may have biased the results towards those practitioners who were more likely to have an interest in AMS, and more awareness of recommended prescribing practices.

102 Establishment of formalized AMS programs has been identified as one of the key strategies for addressing AMR in Australia’s National AMR implementation strategy5 and is critical in providing veterinarians with the knowledge and tools to reduce inappropriate prescribing of antimicrobials in animals. The Australian state and territory veterinary boards should coordinate with government, professional bodies and academic institutions examining the topic, to require AMS in veterinary practices. This study has provided insights into the barriers and enablers for AMS in Australian veterinary practices (Table 2) and has suggested a number of measures that may support the establishment of veterinary AMS programs in Australia (Table 3).

Table 3. Summary of recommendations to facilitate the establishment of AMS programs in veterinary practices. Observed gap Recommendations Veterinary AMS legislation Require veterinary practices to have AMS policies Restrict antimicrobial sales that occur without formal consultation

Education & training Develop online courses and training on AMS targeted at veterinary practitioners (may contribute to continuing education requirements) Provide courses and training on AMS processes to specialists

Resources Develop a means of easily monitoring antimicrobial use and resistance in veterinary practice Develop therapeutic guidelines for antimicrobial use in animals Make available examples and templates for AMS policies and procedures, including templates for on-farm use of antimicrobials

10

3 Supplementary Table 1. Coding tree Final coding framework Initial coding framework Client expectations Client misuse of antimicrobials Convenience of some antimicrobials Fear of something going wrong Clients not concerned about AMR Lack of client education If I don’t, someone else will AMR perceptions Concerned for human health Fear of AMR in the future Fear of losing prescribing rights Education Unsure of significance of AMR Habit & personal experience Cost of continuing education Lack of time In-clinic case review Specialist advice Costs Cost of culture & sensitivity Costs of diagnostic testing Clinic structure Lack of support by clinic structure Colleagues for advice Resources Lack of guidelines Farm protocols AMR in practice Very low levels of resistance encountered Low levels of resistance encountered High levels of resistance encountered Diagnostic tests In-house diagnostics Laboratory tests Antimicrobials with high importance rating Cefovecin Fluoroquinolones Pride in profession Pride in service provided Keep costs down for clients Loosing independence Open to change AMS Processes Should restrict some antimicrobials No champions

References 1. Alexander TW, Inglis GD, Yanke LJ, et al. Farm-to-fork characterization of Escherichia coli associated with feedlot cattle with a known history of antimicrobial use. Int J Food Microbiol 2010;137:40-48. 2. Chantziaras I, Boyen F, Callens B, et al. Correlation between veterinary antimicrobial use and antimicrobial resistance in food-producing animals: a report on seven countries. J Antimicrob Chemother 2014;69:827-834. 3. Davey P, Brown E, Charani E, et al. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev 2013:CD003543. 4. World Health Organisation. Global action plan on antimicrobial resistance. In. http://www.who.int/antimicrobial-resistance/global-action-plan/en/: 2015. 5. Commonwealth of Australia. National antimicrobial resistance strategy 2015-2019. In. http://www.health.gov.au/internet/main/publishing.nsf/Content/1803C433C71415CACA 257C8400121B1F/$File/amr-strategy-2015-2019.pdf: 2016. 6. Australian Bureau of Statistics. 8564.0 - Veterinary services, Australia, 1999-2000. In. www.abs.gov.au: 2001. 7. O'Cathain A, Murphy E, Nicholl J. Three techniques for integrating data in mixed methods studies. BMJ 2010;341:c4587. 8. Morgan DL. Practical strategies for combining qualitative and quantitative methods: applications to health research. Qual Health Res 1998;8:362-376.

10

4 Michie S, van Stralen MM, West R. The behaviour change wheel: a new method for characterising and designing behaviour change interventions. Implement Sci 2011;6:42. 9. Pope C, Ziebland S, Mays N. Qualitative research in health care. Analysing qualitative data. BMJ 2000;320:114-116. 10. Liamputtong P. Making sense of qualitative data: the analysis process. In: Qualitative research methods, 3rd ed. Melbourne: Oxford University Press; 2009:277-296. 11. Morse J. 'Emerging from the data': the cognitive process of analysis in qualitative inquiry. In: Issues in Qualitative Research Methods. Thousand Oaks, CA: Sage; 1994:23-43. 12. Bradley EH, Curry LA, Devers KJ. Qualitative data analysis for health services research: developing taxonomy, themes, and theory. Health Serv Res 2007;42:1758-1772. 13. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology 2006;3:77-101. 14. Australian Veterinary Association. Australian veterinary workforce survey 2014. In. www.ava.com.au/workforce-data: 2014. 15. Australian Strategic and Technical Advisory Group on Antimicrobial Resistance. Importance rating and summary of antibacterials used in human health in Australia. In. 16. http://www.health.gov.au/internet/main/publishing.nsf/Content/ohp-amr.htm: Commonweath of Australia; 2015. Holloway S, Trott DJ, Shipstone M, et al. Antibiotic Prescribing: detailed guidelines. In. www.ava.com.au/sites/default/files/AVA_website/pdfs/AIDAP guidelines.pdf: 17. Australasian Infectious Diseases Advisory Panel; 2013. British Small Animal Veterinary Association. PROTECT. In. .aspx: 2016. 18. McDougall S, Compton C, Botha N. Factors influencing antimicrobial prescribing by https://www.bsava.com/Resources/Veterinary-resources/PROTECTveterinarians and usage by dairy farmers in New Zealand. N Z Vet J 2017;65:84-92. 19. Visschers VH, Postma M, Sjolund M, et al. Higher perceived risk of antimicrobials is related to lower antimicrobial usage among pig farmers in four European countries. Vet Rec 20. 2016;179:490. De Briyne N, Atkinson J, Pokludova L, et al. Factors influencing antibiotic prescribing habits and use of sensitivity testing amongst veterinarians in Europe. Vet Rec 2013;173:475. 21. Dellit TH, Owens RC, McGowen JE, et al. Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America: guidelines for developing an institutional 22. program to enhance antimicrobial stewardship. Clin Infect Dis 2007;44:159-177. Duguid M, Cruickshank M. Antimicrobial stewardship in Australian hospitals. In. Sydney: Australian Commission on Safety and Quality in Health Care; 2011. 23. CDC. Core elements of hospital antibiotic stewardship programs. In. www.cdc.gov/getsmart/healthcare/implementation/core-elements.html: US Department of 24. Health and Human Services; 2014. Public Health England. Start Smart - Then focus: antimicrobial stewardship toolkit for English hospitals. In. www.gov.uk/phe: 2015. 25. Charani E, Edwards R, Sevdalis N, et al. Behavior change strategies to influence antimicrobial prescribing in acute care: a systematic review. Clin Infect Dis 2011;53:651-662. 26. Rawson TM, Charani E, Moore LS, et al. Mapping the decision pathways of acute infection management in secondary care among UK medical physicians: a qualitative study. BMC Med 27. 2016;14:208. Charani E, Castro-Sanchez E, Sevdalis N, et al. Understanding the determinants of antimicrobial prescribing within hospitals: the role of "prescribing etiquette". Clin Infect Dis 28. 2013;57:188-196. Coyne LA, Latham SM, Williams NJ, et al. Understanding the culture of antimicrobial prescribing in agriculture: a qualitative study of UK pig veterinary surgeons. J Antimicrob 29. Chemother 2016;71:3300-3312.

10

5 30. Fletcher-Lartey S, Yee M, Gaarslev C, et al. Why do general practitioners prescribe antibiotics for upper respiratory tract infections to meet patient expectations: a mixed methods study. BMJ Open 2016;6:e012244.

10

6 BACK TO TABLE OF CONTENTS

Chapter 9:

THE ROLE OF VETERINARY DIAGNOSTIC LABORATORIES IN PROMOTING ANTIMICROBIAL STEWARDSHIP

10

7 Antimicrobial susceptibility testing methods used by the Australian Veterinary Diagnostic Laboratories

LY Hardefeldta,b, M Marendaa, H Crabba,b, MA Stevensona, JR Gilkersona, H Billman- Jacobea,b, and GF Browninga,b aAsia-Pacific Centre for Animal Health, Melbourne Veterinary School, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, Victoria 3010, Australia, and 250 Princes Highway, Werribee, Victoria 3030, Australia bNational Centre for Antimicrobial Stewardship, Peter Doherty Institute, Grattan St, Parkville, Victoria Australia

This paper has been accepted for publication by the Australian Veterinary Journal.

Abstract The national strategy for tackling antimicrobial resistance highlights the need for antimicrobial stewardship in veterinary practice and for surveillance of antimicrobial susceptibility in veterinary pathogens. Diagnostic laboratories have an important role in facilitating both of these processes, but it is not clear whether data from veterinary diagnostic laboratories is similar enough to allow for compilation and if there is consistent promotion of appropriate antimicrobial use embedded in the approaches of different laboratories to susceptibility testing. A cross-sectional study of antimicrobial susceptibility testing and reporting procedures by Australian veterinary diagnostic laboratories was conducted in 2017 using an online questionnaire. All veterinary diagnostic laboratories in Australia completed the questionnaire. Kirby-Bauer disc diffusion was the method predominantly used for antimicrobial susceptibility testing and was used to evaluate 86% of all isolates, although two different protocols were used across the 18 laboratories (CLSI 15/18, CDS 3/18). Minimum inhibitory concentrations were never reported by 61% of laboratories. Common isolates were consistently reported on across all species, except for Gram negative isolates in pigs, for which there was some variation in the approach to reporting. There was considerable diversity in the panels of antimicrobials used for susceptibility testing on common isolates, and no consistency was apparent between laboratories for any bacterial species. We recommend that that nationally agreed and consistent antimicrobial panels for routine susceptibility testing should be developed and a uniform set of guidelines should be adopted by veterinary diagnostic laboratories in Australia.

Abbreviations AMR, antimicrobial resistance; CDS, calibrated dichotomous sensitivity; CLSI, Clinical Laboratory Standards Institute; EUCAST, European Committee on Antimicrobial Susceptibility Testing

Introduction Antimicrobial resistance (AMR) of animal origin has been implicated in the development of AMR in humans1-6. Both global7 and national strategies8 for tackling antimicrobial resistance have called for improved antimicrobial stewardship in animal health and for the monitoring of antimicrobial susceptibility patterns. However, surveillance of resistant veterinary pathogens is limited in Australia because of a lack of resources and heavy reliance on passive surveillance of clinical specimens. Moreover, comparative studies require strict quality control procedures to ensure data are

10

8 uniform. It is unclear if data from veterinary diagnostic laboratories is similar enough to allow national compilation.

Antimicrobial stewardship is a term used to describe the implementation of processes to ensure the appropriate use of antimicrobials in animals and people, with the goal of reducing the driving factor (antimicrobial use) for AMR.8 Many antimicrobial stewardship programs call for an increase in the use of culture and susceptibility testing, to improve appropriate antimicrobial use in medical9 and veterinary10 clinical practice.

Previous research has shown that antimicrobial susceptibility results obtained from veterinary diagnostic laboratories are rarely scrutinised by veterinarians in clinical veterinary practice (Hardefeldt et.al, in review). In addition, reporting of sensitivity patterns may lead to changes in antimicrobial selection by veterinarians in practice, so there is a need for these to be consistent between diagnostic laboratories. Variability in testing methodology and data storage have been identified as barriers to the use of veterinary diagnostic laboratory data for AMR surveillance in the United States.11

The aim of this study was to investigate laboratory procedures that may contribute to antimicrobial selection by veterinarians in Australia and to evaluate whether a lack of consistency in procedures could affect the compilation of data that may contribute to an effective AMR surveillance system.

Methods A cross-sectional study of antimicrobial susceptibility testing and reporting procedures by Australian veterinary diagnostic laboratories was conducted in 2017. The source population consisted of the 18 veterinary diagnostic laboratory organisations providing services for veterinary clinical specimens in Australia in 2017.

A letter of invitation was sent to a microbiologist at each laboratory. Questionnaire responses were collected and managed using REDCap electronic data capture tools.12 The questionnaire was comprised of five sections (available as supplementary material). The initial section asked for information on the frequency with which different methods of assessing antimicrobial susceptibility were used and the standards that were used to perform this testing (CLSI, CDS, EUCAST). The second section asked which organisms were never tested for susceptibility. Section three asked respondents to nominate the three most frequently isolated pathogens, for dogs, cats, horses, cattle and pigs, and to nominate the antimicrobials that were routinely included in susceptibility testing performed on these pathogens. Section four asked about the approach used by the laboratory to identify methicillin-resistant Staphylococcus species, to detect extended spectrum beta-lactamase producing Enterobacteriaceae and to detect vancomycin resistant Enterococcus species. Finally, respondents were asked to nominate if and, if so, when antimicrobial susceptibility testing was performed but not reported to the clinician submitting the request for diagnostic services. The survey was pre-tested with two independent laboratory personnel. The survey took 15-20 minutes to complete. Minimal changes were made to the survey following the initial testing.

Data were downloaded from the survey software to spreadsheets and processed with Microsoft Office Excel, 2016. The respondents had to indicate that they routinely

10

9 performed culture and susceptibility testing on each species to be included in the analysis for that species. Descriptive statistics were computed, with percentages being reported as a proportion of the total laboratories responding to each animal species section.

This research was approved by the University of Melbourne Faculty of Veterinary and Agricultural Science Human Ethics Advisory Group under Approval No. 1749375.2.

Results All 18 of the veterinary diagnostic laboratory organisations operating in Australia in 2017 were recruited into this study and all completed the survey. Of these, 12 routinely processed dog and cat isolates, 13 laboratories routinely processed equine isolates, 14 laboratories routinely processed bovine isolates and 9 laboratories routinely processed porcine isolates.

The Kirby-Bauer disc diffusion method was the method predominantly used for susceptibility testing. The disc diffusion method was used by all laboratories, with an average across the laboratories of 86% of all isolates tested using this method. Broth dilution methods were used by eight laboratories, with an average across these laboratories of 8% of all isolates tested using this method. The Etest® method was used by 5 laboratories, with an average across the laboratories of 6% of all isolates tested using this method. Agar dilution methods were used by 3 laboratories, with an average across the laboratories of less than 1% of all isolates tested using this method. Genotyping for antimicrobial resistance genes was rarely used, with a detection assay for the mecA gene in Staphylococcus isolates only offered if requested by a clinician or researcher. Laboratories were typically reporting the susceptible-intermediate- resistant interpretation for disc diffusion results (14/18), with others reporting zone diameters (4/18). Minimum inhibitory concentrations (MICs) were never routinely reported by any laboratory and were not reported at all by most laboratories (11/18), with the remainder reporting MICs only when they were requested by clinicians (7/18). The Clinical and Laboratory Standards Institute (CLSI) guidelines were used by 83% (15/18) of laboratories, with the remainder using the Calibrated Dichotomous Sensitivity (CDS) test method. Reasons for not performing susceptibility testing included isolation of fastidious organisms (8/18), anaerobic organisms (7/18), and organisms with predictable susceptibility patterns (3/18).

The most commonly isolated pathogens for each species were relatively consistent between laboratories for all species, except for pigs, where there was diversity, predominantly among gram negative isolates (Table 1). One laboratory did not report Gram positive equine isolates, so this laboratory was excluded for that section.

To further analyse the patterns of antimicrobial susceptibility testing, the data were compared when more than half of the laboratories were routinely culturing and testing isolates of a specific bacterial species. There was considerable diversity in the panels of antimicrobials used in susceptibility testing for the common isolates, with no consistency detected between laboratories in any animal or bacterial species (Table 2). Canine Streptococcus spp. isolates were tested for amoxycillin, ampicillin or penicillin susceptibility by 64% (7/11) of laboratories reporting these isolates. Equine and bovine Streptococcus spp. isolates were frequently tested for susceptibility to3rd generation

1

10 Table 1. Most commonly isolated pathogens for each animal species included in the survey Canine Feline Equine Bovine Porcine Gram positive Staphylococcus spp. 12/12 12/12 11/12 12/14 6/9 Streptococcus spp . 11/12 7/12 12/12 13/14 7/9 Enterococcus spp. 9/12 9/12 3/12 Corynebacterium spp. 1/12 1/12 2/12 2/14 Clostridium spp. 1/12 1/14 Rhodococcus spp. 1/12 Trueperella pyogenes 5/14 5/9 Erysipelothrix spp. 2/9 Gram negative Pseudomonas spp. 12/12 6/12 9/13 1/14 1/9 Escherichia coli 12/12 12/12 11/13 11/14 5/9 Proteus spp. 9/12 6/12 1/13 Pasteurella spp. 2/12 4/12 6/14 4/9 Other Enterobacteriaceae 1/12 1/12 1/13 2/14 Serratia spp. 1/12 Salmonella spp. 1/12 5/13 6/14 1/9 Klebsiella spp. 1/12 1/13 1/9 Actinobacillus spp. 2/13 4/9 Brucella spp. 1/13 Mannheimia haemolytica 3/14 Enterobacter spp. 2/14 1/9 Yersinia spp. 1/14 Moraxella bovis 1/14 Haemophilus spp. 1/9 No. laboratories reporting that these bacterial species were among the three most common species isolated from this animal species/No. laboratories that commonly receive samples from this animal species cephalosporins (11/12 laboratories and 7/14 laboratories, respectively). Susceptibility testing for vancomycin was rare for all bacterial isolates and animal species, except for Enterococcus isolates from dogs and cats.

Susceptibility testing for antimicrobials that are not permitted to be used in food producing animals was uncommon, and when susceptibility testing was performed it was for identification purposes or to monitor for resistance, and was not reported to clinicians. Other reasons for not reporting susceptibility data included when the susceptibility test was performed for research purposes.

Routine susceptibility testing for antimicrobials with a high importance rating (as defined by the Australian Strategic Advisory Group on Antimicrobial Resistance13) was generally uncommon, except for fluoroquinolones, 3rd generation cephalosporins and amikacin. Susceptibility to imipenem was infrequently reported in any species. Testing for amikacin, piperacillin and ticarcillin-clavulanate was uncommon in dog and cat isolates. Sensitivity to vancomycin was rarely assessed, except for Enterococcus spp. isolates. Ticarcillin-clavulanate was only reported for E. coli in dogs (5/12) and cats (5/12) and for Pseudomonas isolates in dogs (6/12). Fluoroquinolone susceptibility was commonly assessed for Staphylococcus spp. (12/12), E. coli (11/12), Pseudomonas spp. (11/12) and Proteus spp. (9/9) isolates in laboratories frequently isolating these pathogens from dogs and for Staphylococcus spp. (11/12), Streptococcus spp. (6/7), Enterococcus spp. (7/9), E. coli (10/12), Pseudomonas spp. (6/6) and Proteus spp. (6/6) isolates in laboratories frequently isolating these pathogens from cats. Susceptibility to

1

11 Table 2. Antimicrobials commonly used included in susceptibility testing for different pathogens Species Isolate Antimicrobials included in susceptibility testing by >90% of laboratories Canine Staphylococcus spp. Amoxycillin-clavulanate Clindamycin Fluoroquinolones Tetracycline Trimethoprim-sulphonamide Streptococcus spp. Tetracyclines Enterococcus spp. Ampicillin or amoxycillin Tetracyclines E. coli Ampicillin or amoxycillin Amoxycillin-clavulanate Fluoroquinolones Gentamicin Tetracyclines Trimethoprim-sulphonamide Pseudomonas spp. 3rd generation cephalosporins Fluoroquinolones Gentamicin Proteus spp. Ampicillin or amoxycillin Amoxycillin-clavulanate Fluoroquinolones Gentamicin Trimethoprim-sulphonamide

Feline Staphylococcus spp. Amoxycillin-clavulanate Clindamycin Fluoroquinolones Tetracycline Trimethoprim-sulphonamide Streptococcus spp. Fluoroquinolones Tetracyclines Trimethoprim-sulphonamide Enterococcus spp. Ampicillin or amoxycillin Amoxycillin-clavulanate Fluoroquinolones Tetracyclines E. coli Ampicillin or amoxycillin Amoxycillin-clavulanate Fluoroquinolones Gentamicin Tetracyclines Trimethoprim-sulphonamide Pseudomonas spp. Fluoroquinolones Proteus spp. 3rd generation cephalosporins Ampicillin or amoxycillin Amoxycillin-clavulanate Fluoroquinolones Gentamicin Tetracyclines Trimethoprim-sulphonamide

Equine Streptococcus spp. 3rd generation cephalosporins Penicillin

1

12 Tetracycline Trimethoprim-sulphonamide Staphylococcus spp. 3rd generation cephalosporins Gentamicin Penicillin Tetracycline Trimethoprim-sulphonamide E. coli 3rd generation cephalosporins Enrofloxacin Gentamicin Tetracycline Trimethoprim-sulphonamide Pseudomonas spp. Enrofloxacin Gentamicin Trimethoprim-sulphonamide

Bovine Staphylococcus spp. Penicillin Tetracycline Trimethoprim-sulphonamide Streptococcus spp. Penicillin Tetracycline Pasteurella spp. 3rd generation cephalosporins Amoxycillin or ampicillin Neomycin Tetracycline Trimethoprim sulphonamide E. coli Amoxycillin or ampicillin Neomycin Tetracycline Trimethoprim-sulphonamide Salmonella spp. Amoxycillin or ampicillin Neomycin Tetracycline Trimethoprim-sulphonamide

Porcine Streptococcus spp. Penicillin Tetracycline Staphylococcus spp. Penicillin Tetracycline Trimethoprim-sulphonamide Trueperella pyogenes Penicillin Tetracycline Trimethoprim-sulphonamide E. coli Amoxycillin or ampicillin Tetracycline Trimethoprim-sulphonamide

3rd generation cephalosporins was commonly assessed for Pseudomonas spp. isolates (10/12), Staphylococcus spp. (9/12), E. coli (9/12), Proteus spp. (7/9) and Streptococcus spp. (4/11) in laboratories frequently isolating these pathogens from dogs, and Staphylococcus spp. (7/12), Streptococcus spp. (6/7), Enterococcus spp. (4/9), E. coli (8/12), Pseudomonas spp. (3/6) and Proteus spp. (5/6) isolates in laboratories frequently isolating these pathogens from cats. Piperacillin susceptibility was never tested for in equine isolates and testing for vancomycin, imipenem and ticarcillin- clavulanate susceptibility was reported to be rare. Susceptibility to 3rd generation

11

3 cephalosporins was commonly assessed for Streptococcus spp. (11/12), Staphylococcus spp. (9/11), E. coli (11/13) and Pseudomonas spp. (5/9) isolates in laboratories frequently isolating these pathogens from horses. Fluoroquinolone susceptibility testing was also common for Streptococcus spp. (8/12), Staphylococcus spp. (8/11), E. coli (10/11) and Pseudomonas spp. (7/9) isolates from horses. Amikacin susceptibility testing was commonly reported for Streptococcus spp. (5/12), Staphylococcus spp. (6/11), E. coli (8/11) and Pseudomonas spp. (5/9) isolates from horses. In cattle, the only antimicrobials with a high importance rating for which susceptibility were reported were 3rd generation cephalosporins, for Staphylococcus spp. (5/12), Streptococcus spp. (7/13), E. coli (6/11), Pasteurella spp. (4/6), and Salmonella spp. (2/6). Similarly, for pig isolates 3rd generation cephalosporins were the only antimicrobials with a high importance rating for which susceptibility was reported and this was infrequently and only for Staphylococcus spp. (2/6), E. coli (1/5) and Trueperella pyogenes (2/5).

Discussion This is the first survey to investigate antimicrobial susceptibility testing procedures of Australian veterinary diagnostic laboratories and has quantified the variety of testing procedures and antimicrobial panels that are currently being used. Veterinary diagnostic laboratories have been identified internationally as a vital source of data for monitoring antimicrobial susceptibility,11 and the Office International des Epizooties (OIE) Ad Hoc Group of Experts on Antimicrobial Resistance has called for standardisation of susceptibility testing and harmonisation of the antimicrobials used in susceptibility testing to enable international comparison of data.14 Standardisation is also required to allow for national compilation of data. However, there are limitations to the use these data for surveillance, as clinical isolates might be expected to have higher levels of resistance resulting from the selective effect of recent exposure to antimicrobials, than isolates collected from sentinel animals.

This survey has shown that there is no uniformity in the practice of susceptibility testing, with two laboratory standards being used commonly (CLSI and CDS), and a range of other methods employed for specific circumstances. A number of different susceptibility testing methods are also used in the human medical laboratory sector.15 While a national approval system for laboratories does not exist, the National Association of Testing Authorities, Australia (NATA) accredits veterinary diagnostic laboratories for a range of services. Nine of the 18 veterinary laboratories are accredited for antimicrobial susceptibility testing (to the standard ISO/IEC 17025:2005:20.11.02).16 NATA accreditation requires participation in regular proficiency testing. Non-accredited laboratories may also participate in these programs, although the extent to which the laboratories that are not accredited for antimicrobial testing participate in proficiency testing is not known.

Ideally, a national surveillance system would be based on MICs, as this allows for subtle shifts in susceptibility to be detected, rather than a dichotomous system of sensitive or resistant. Of concern is that this study has shown that many (11 of 18) veterinary diagnostic laboratories currently do not perform any MIC testing. This is a significant barrier that will be difficult to overcome, given the costs involved in acquiring appropriate equipment and the barrier that the cost of culture and susceptibility testing to animal owners poses to greater utilisation of susceptibility testing (Hardefeldt et.al.,

11

4 in review). An additional barrier is that of susceptibility breakpoints. There are relatively few validated veterinary breakpoints and many interpretations of susceptibility are based on extrapolation from human breakpoints. While the CLSI guidelines recommend that the laboratory should inform veterinarians of the source of the breakpoints, this is rarely done. In addition, there is substantial variation between the breakpoints set by CDS and CLSI, making comparison, and interpretation, difficult. Substantial variability was also found in the range of antimicrobials used in susceptibility testing for individual bacterial species. Consensus on standard panels is clearly needed.

Veterinary diagnostic laboratories also have a significant role in antimicrobial stewardship, as veterinarians rely on susceptibility reports to guide antimicrobial selection. Previous research has identified that Australian veterinarians rarely consulted laboratory microbiologists about culture and susceptibility results, even when these results were unexpected (Hardefeldt et. al., in review). In this study, testing and reporting of susceptibility to three antimicrobials with a high-importance rating was common - for 3rd generation cephalosporins in all species, fluoroquinolones in dogs, cats and horses, and amikacin in horses. Beta-haemolytic Streptococcus spp. isolates were the most commonly reported Streptococcus spp. isolates for dogs and horses, and these are uniformly sensitive to penicillins.17, 18 Reporting of 3rd generation cephalosporin susceptibility for beta-haemolytic Streptococcus spp. isolates in dogs and horses may lead practitioners to choose these drugs over antimicrobials with a low- importance rating, such as amoxycillin or penicillin. Injudicious antimicrobial susceptibility reporting has resulted in inappropriate use of antimicrobials in human medicine.19 Selective reporting has been used safely in a human hospital,20 and is recommended by EUCAST.21 A similar selective approach to reporting should be implemented in veterinary medicine.

Testing for susceptibility to fluoroquinolones was common for horse, dog and cat isolates, particularly for Gram negative isolates and Staphylococcus species. This may reflect the widespread use of fluoroquinolones in general practice in Australia,22 but may also be encouraging this use. Testing for susceptibility to amikacin was common for equine isolates. Amikacin should be reserved for cases of documented gentamicin resistance, which appears to be uncommon in Australia.23 A further concern this survey has raised is the low frequency with which many laboratories included appropriate narrow spectrum agents in their susceptibility panels. Pasteurella spp. are commonly susceptible to penicillin, but the antimicrobials most commonly included in testing panels for Pasteurella spp. were the broad-spectrum antimicrobials, trimethoprim- sulphonamide and tetracycline, and antimicrobials with high-importance rating, 3rd generation cephalosporins, while only 50% of laboratories included penicillin in their panels for these species. Susceptibility to antimicrobials with a high-importance rating should only be reported if the organism is resistant to all agents with a low- or medium- importance rating, or if specifically requested by clinicians. In addition, narrow- spectrum antimicrobials should be preferred over broad-spectrum antimicrobials.

Some surprising pathogens were listed among the three most commonly isolated organisms in some species - for example Brucella spp. in horses. This may reflect the low number of samples subjected to susceptibility testing by some laboratories serving the equine sector, or may reflect regional differences in equine diseases.

11

5 In conclusion, this survey has identified several barriers to the use of veterinary diagnostic laboratory data for antimicrobial resistance surveillance and antimicrobial stewardship. Standardisation of laboratory methods should be considered and standard panels of antimicrobials for susceptibility testing for common isolates need to be generated and adopted.

References 1. Platell JL, Cobbold RN, Johnson JR et al. Commonality among fluoroquinolone-resistant sequence type ST131 extraintestinal Escherichia coli isolates from humans and companion animals in Australia. Antimicrob Agents Chemother 2011;55:3782-3787. 2. Liu W, Liu Z, Yao Z et al. The prevalence and influencing factors of methicillin-resistant Staphylococcus aureus carriage in people in contact with livestock: a systematic review. Am J Infect Control 2015;43:469-475. 3. Dohmen W, Bonten MJ, Bos ME et al. Carriage of extended-spectrum beta-lactamases in pig farmers is associated with occurrence in pigs. Clin Microbiol Infect 2015;21:917-923. 4. Schwaber MJ, Navon-Venezia S, Masarwa S et al. Clonal transmission of a rare methicillin- resistant Staphylococcus aureus genotype between horses and staff at a veterinary teaching hospital. Vet Microbiol 2013;162:907-911. 5. Paul NC, Moodley A, Ghibaudo G, Guardabassi L. Carriage of methicillin-resistant Staphylococcus pseudintermedius in small animal veterinarians: indirect evidence of zoonotic transmission. Zoonoses Public Health 2011;58:533-539. 6. Boost M, Ho J, Guardabassi L, O'Donoghue M. Colonization of butchers with livestock- associated methicillin-resistant Staphylococcus aureus. Zoonoses Public Health 2013;60:572-576. 7. World Health Organisation. Global action plan on antimicrobial resistance. http://www.who.int/antimicrobial-resistance/global-action-plan/en/, 2015. 8. Commonwealth of Australia. National antimicrobial resistance strategy 2015-2019. http://www.health.gov.au/internet/main/publishing.nsf/Content/1803C433C71415CAC A257C8400121B1F/$File/amr-strategy-2015-2019.pdf, 2016. 9. Australian Commission on Safety and Quality in Health Care. Antimicrobial stewardship in Australian hospitals. Australian Commission on Safety and Quality in Health Care,, https://www.safetyandquality.gov.au/wp-content/uploads/2011/01/Antimicrobial- stewardship-in-Australian-Hospitals-2011.pdf, 2011. 10. Guardabassi L, Prescott JF. Antimicrobial stewardship in small animal veterinary practice: from theory to practice. Vet Clin North Am Small Anim Pract 2015;45:361-376, vii. 11. Brooks MB, Morley PS, Dargatz DA et al. Survey of antimicrobial susceptibility testing practices of veterinary diagnostic laboratories in the United States. J Am Vet Med Assoc 2003;222:168-173. 12. Harris PA, Taylor R, Thielke R et al. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377-381. 13. Australian Strategic and Technical Advisory Group on Antimicrobial Resistance. Importance rating and summary of antibacterials used in human health in Australia. Commonweath of Australia, http://www.health.gov.au/internet/main/publishing.nsf/Content/ohp-amr.htm, 2015. 14. Franklin A, Acar J, Anthony F et al. Antimicrobial resistance: harmonisation of national antimicrobial resistance monitoring and surveillance programmes in animals and in animal-derived food. Rev Sci Tech 2001;20:859-870. 15. Merlino J. Antibiotic susceptibility testing methods and emerging bacterial resistance in hospitals. Microbiology Australia 2014;35:9-12. 16. National Association of Testing Authorities A. Veterinary facilities list. 2017. Retrieved 13/10/17.

11

6 17. Windahl U, Holst BS, Nyman A, Gronlund U, Bengtsson B. Characterisation of bacterial growth and antimicrobial susceptibility patterns in canine urinary tract infections. BMC Vet Res 2014;10:217. 18. Moore RM, Schneider RK, Kowalski J et al. Antimicrobial susceptibility of bacterial isolates from 233 horses with musculoskeletal infection during 1979-1989. Equine Vet J 1992;24:450-456. 19. Steffee CH, Morrell RM, Wasilauskas BL. Clinical use of rifampicin during routine reporting of rifampicin susceptibilities: a lesson in selective reporting of antimicrobial susceptibility data. J Antimicrob Chemother 1997;40:595-598. 20. Cunney R, Aziz HA, Schubert D, McNamara E, Smyth E. Interpretative reporting and selective antimicrobial susceptibility release in non-critical microbiology results. J Antimicrob Chemother 2000;45:705-708. 21. Leclercq R, Canton R, Brown DF et al. EUCAST expert rules in antimicrobial susceptibility testing. Clin Microbiol Infect 2013;19:141-160. 22. Hardefeldt LY, Holloway S, Trott DJ et al. Antimicrobial prescribing in dogs and cats in Australia: results of the Australasian Infectious Disease Advisory Panel survey. J Vet Intern Med 2017;31:1100-1107. 23. Russell CM, Axon JE, Blishen A, Begg AP. Blood culture isolates and antimicrobial sensitivities from 427 critically ill neonatal foals. Aust Vet J 2008;86:266-271.

11

7 BACK TO TABLE OF CONTENTS

Chapter 10:

General Discussion

11

8 It is clear that AMR is a global health challenge. While the extent of the contribution to community AMR cannot currently be elucidated, it is clear that antimicrobial resistance does develop as a result of antimicrobial use in animals and that this does pose a risk to people, especially those in close contact with animals, and the families of these people. However, antimicrobial use in animals can be justified, on grounds of animal welfare, food security and food safety, but inappropriate use of antimicrobials cannot be justified.

The overall aims of this body of research were to investigate rates of antimicrobial use in companion animals in Australia; to investigate detailed antimicrobial usage to identify causes of inappropriate antimicrobial use in companion animals, horses and cattle; to identify the enablers to, and barriers of, antimicrobial stewardship in veterinary practices in Australia; and finally to develop an economically viable, effective and adaptable antimicrobial stewardship package for Australian veterinary practices.

Assessing rates of antimicrobial use in veterinary medicine Companion animals live intimately with many families1. Companion animals can be reservoirs of AMR organisms that can infect people, and can also be infected by these organisms from human source, compromising animal welfare2-6. As antimicrobial drug use exerts a selective pressure on AMR bacteria, surveillance of antimicrobial use in companion animals is critical. This thesis documented the rates of antimicrobial exposure in companion animals in the Australian community in chapter 1. These rates are lower than in the human community, both in Australia and in similar developed nations. However, as this is the first study to examine the rates of antimicrobial exposure in a population of companion animals, global comparisons across animal populations are not possible. The data source that made this research possible is unique to small companion animals in Australia, as full medical insurance for horses is rare and is not available for production animal species. Other means of collecting large cohort data, independent of veterinary intervention, is expensive, time-consuming and therefore is unlikely to be successful. In fact, even collecting prescribing data on animals that have had veterinary intervention is difficult. Veterinarians in Australia are predominately based in small businesses, with limited resources for additional administration and limited time for activities, such as antimicrobial use reporting, that do not generate income.

VetCompassÒ is another tool that may provide a means of accessing prescribing data, and has been used for some studies in the UK7, 8. However, VetCompassÒ is restricted to collecting data on companion animal species, and therefore will not be a solution to the current problem of collecting data from a variety of species, although similar programs have been able to collect prescribing data from equine patients9. A flexible tool for monitoring antimicrobial use has been designed and methods for delivering this to both intensive animal veterinarians and to veterinary practices are currently being explored. Methods that have been trialled include having veterinary student collect data while visiting practices in their final years of study, and having veterinary practitioners collect data. However, both of these methods require substantial time commitments, and neither have been

11

9 successful in collecting large scale data. As small businesses shift to digital management environments there may be an opportunity to automate this surveillance tool in a way that is acceptable to veterinary practice owners. Such methods are currently undergoing evaluation.

Causes of inappropriate antimicrobial use by veterinarians Until these tools are developed, analysis of detailed antimicrobial use is often restricted to the methods used in the other studies described in this thesis, surveys of antimicrobial use for hypothetical clinical scenarios. This method allowed the investigation of detailed antimicrobial use in these studies and identified key areas where antimicrobials are used inappropriately. Across all sectors of veterinary practice there was low use of antimicrobials with a high importance rating, with the exception of the 3rd generation cephalosporins. Registered products from this class are available in all animal health sectors and appear to be used frequently, especially where long-acting formulations exist, and in the dairy sector because milk does not need to be withheld from sale after use. In Europe, the use of critically important antimicrobials appears higher10, although direct comparisons are misleading as 4th generation cephalosporins are registered in Europe, and not in Australia, and fluoroquinolones have wider registration than in Australia. The use of these agents for convenience, rather than because of a need for their specific spectrum of activity, is concerning, especially in dairy cattle practice, where extended spectrum cephalosporin resistance in Salmonella isolates is an emerging issue11. Anecdotally, use also appears to be widespread in the equine sector, and the consequences in this species may be masked by the limited AMR surveillance undertaken in horses.

Under-dosing of antimicrobial agents was identified as an issue in prescribing for both cattle and horses. Inappropriate drug labelling is likely to be contributing to this. This is of particular concern in food producing animals, as deriving an appropriate withholding period for antimicrobials administered at a dose higher than that on the label is very difficult. However, the labels on older companion animal medications, such as amoxycillin and amoxycillin/clavulanate, are also inappropriate and yet the vast majority of companion animal veterinarians are using appropriate doses. Recent graduates in companion animal practice were less likely to be compliant with guidelines, and a similar trend was seen in equine veterinarians, which may reflect the reliance of these groups on labels and the advice of colleagues, as has been found elsewhere12, 13, for guidance on drug dosing, especially when in an ambulatory setting, as is typically the case in equine and bovine practice in Australia. In addition, most of the respondents to the equine and bovine surveys were mixed practitioners, and thus may be less familiar with the full range of recommended doses, compared to the respondents to the companion animal survey, who were predominately only participating in companion animal practice. While this factor did not reach significance in a multivariable model, the survey was not designed nor powered to investigate this, and further investigation would be needed to fully elucidate this risk factor for suboptimal prescribing. However, it seems reasonable that veterinarians prescribing across multiple sectors would have less continuing education on each specific species that they work with than veterinarians working in only one area of

1

20 practice. Consistent with this, discussion of clinical cases with peers and clinic meetings have previously been identified as being useful forums for veterinarians to share knowledge13. A freely and readily available source of reliable information on appropriate antimicrobial prescribing is clearly needed for Australian veterinarians.

Enablers and barriers to implementing antimicrobial stewardship The first part of this thesis identified a clear need for antimicrobial stewardship measures to be developed for veterinary practices in Australia. Antimicrobial stewardship measures implemented in hospitals have been successful for three main reasons. Firstly, legislation often requires hospitals to have AMS measures in place; secondly, there is typically support from the governing body overseeing the hospital, allowing financial resources to be allocated to such programs; and finally, AMS is often associated with lower costs for antimicrobials and hence results in a net financial benefit to the hospital budget. These enablers are not transferable to veterinary practice. There is no requirement for veterinary AMS, veterinarians are typically small businesses, and antimicrobial sales represent an income for veterinary practices, not a cost, and therefore this may be a disincentive to fully engage in AMS. These enabling factors and barriers to AMS were investigated in this thesis. Lack of governance was clearly identified as a barrier, but lack of finances and loss of profit from antimicrobial drug sales were not identified as inhibitors. Rather, lack of resources was a significant issue. Antimicrobial use guidelines, as a single intervention, have been successful in reducing antimicrobial use in one referral veterinary hospital in the USA14. This work needs to be replicated in other environments to fully evaluate the effectiveness of such an intervention. However, other AMS interventions have not been investigated in veterinary medicine and this is a substantial gap in our knowledge, as is the lack of evaluation of AMS in general practice veterinary medicine.

Governance is a double-edged sword in veterinary AMS. Competition between practices was identified as an important factor in veterinary prescribing and is very likely to be affecting the rate at which AMS can be implemented. In equine and bovine practice, loss of a single farm as a client can substantially affect the profitability of a veterinary clinic, as some farms contain large numbers of animals and one or more veterinarians may be employed to serve that farm alone. A change in practice policy that restricts antimicrobial use or sales to a farmer may result in that farmer taking their business to a different veterinary clinic and hence affecting the viability of the veterinary practice itself. In this situation, it is clear why veterinary practices would prefer legislation to ensure an even playing field for all competing clinics. However, discussions of governance in veterinary AMS are often directed at dispensing rights and off-label use of antimicrobials. As elucidated is the studies described in this thesis, removal of off-label antimicrobial prescribing rights in veterinary practice would not only be detrimental to AMS, but would also risk the health and welfare of animals under veterinary care. This has recently been highlighted in Europe, where the European Medicines Agency has released reflection paper on off-label use of antimicrobials in animals that is currently open for public comment15. This document discusses many of the issues associated with restricting off-label use of antimicrobials for veterinarians, and these problems

1

21 also are applicable to Australian veterinarians. However, restricting the use of some critically important antimicrobials to their labelled indications may be necessary to curb inappropriate use. This approach has been taken in the USA, where off-label use of 3rd generation cephalosporins and fluoroquinolones is prohibited in major food-producing species16.

Australia is a sparsely populated country with an area of 7.7 million square kilometres and a population density of just 3.1 people per square kilometre17. In a sparsely populated country like Australia, loss of dispensing rights would have a very considerable adverse impact on farmers in rural areas who rely on veterinarians to prescribe and supply antimicrobials 24 hours a day and seven days a week. Decoupling prescribing and dispensing rights was discussed, but eventually rejected, by the European Parliament in 201118. The veterinary profession has an opportunity to lead the implementation of suitable AMS policies that are adaptable to local needs.

An interesting finding in the qualitative survey performed during these studies was that veterinarians rarely consulted with a microbiologist to discuss culture and susceptibility results, even when these results were unexpected. It can be inferred that the veterinary diagnostic laboratories have the potential to play an important role in AMS in the reporting of susceptibility testing. However, the approach adopted by diagnostic laboratories to susceptibility testing and reporting had not previously been investigated in Australia and the survey of the laboratories revealed additional barriers to AMS that need to be addressed. The development of recommended panel of antimicrobials for susceptibility testing, and selective reporting of these panels, for specific organisms in different species should be a priority. Agreement on a laboratory method will also improve the reliability and improve the usefulness of laboratory results in surveillance.

Resource development With gaps in resource availability identified as a barrier - namely availability of guidelines for antimicrobial use and antimicrobial stewardship - steps were taken to address this void. Australian veterinary prescribing guidelines were developed (www.fvas.unimelb.edu.au/vetantibiotics). An emphasis was placed on having evidence-based, independent and evolving guidelines. Peer-reviewed research findings were, and will continue to be, considered and critiqued, and guidelines adjusted if there is an appropriate level of confidence in the research supporting a change. Where evidence is lacking, this is stated and recommendations are then based on human literature and expert opinion. In addition, reference material (a poster for dogs and cats, and a booklet for cattle and horses) was also developed to allow easy access to an abbreviated form of this information in a practice setting (Appendix 1). Validation of these guidelines, and assessment of their implementability, is underway using similar strategies to those that have been used in veterinary medicine in the past and by adapting tools designed for this purpose in human medicine (AGREE II19, GLIA20).

In addition to guidelines on antimicrobial use, guidelines for antimicrobial stewardship were also lacking for veterinary practices in Australia. A generic

1

22 AMS policy was developed as well as an adaptable procedure document (Appendix 2). The policy and procedure provide veterinary practices a comprehensive, yet flexible, antimicrobial stewardship program that can be easily implemented and addresses key areas identified by the Australian National Strategy for Addressing AMR. Research plans have been developed to test the impact of the guidelines on the appropriateness of antimicrobial prescribing in Australian veterinary practices.

References Turner WG. The role of companion animals throughout the family life cycle. Journal of Family Social Work 2005;9:11-21. 1. Guardabassi L, Schwarz S, Lloyd DH. Pet animals as reservoirs of antimicrobial- resistant bacteria. J Antimicrob Chemother 2004;54:321-332. 2. Pantosti A. Methicillin-resistant Staphylococcus aureus associated with animals and its relevance to human health. Front Microbiol 2012;3:127. 3. Couto N, Monchique C, Belas A et al. Trends and molecular mechanisms of antimicrobial resistance in clinical Staphylococci isolated from companion animals 4. over a 16 year period. J Antimicrob Chemother 2016;71:1479-1487. Schwaber MJ, Navon-Venezia S, Masarwa S et al. Clonal transmission of a rare methicillin-resistant Staphylococcus aureus genotype between horses and staff at a 5. veterinary teaching hospital. Vet Microbiol 2013;162:907-911. Paul NC, Moodley A, Ghibaudo G, Guardabassi L. Carriage of methicillin-resistant Staphylococcus pseudintermedius in small animal veterinarians: indirect evidence of 6. zoonotic transmission. Zoonoses Public Health 2011;58:533-539. Buckland EL, O'Neill D, Summers J et al. Characterisation of antimicrobial usage in cats and dogs attending UK primary care companion animal veterinary practices. Vet 7. Rec 2016;179:489. Summers JF, Hendricks A, Brodbelt DC. Prescribing practices of primary-care veterinary practitioners in dogs diagnosed with bacterial pyoderma. BMC Vet Res 8. 2014;10:240. Muellner P, Muellner U, Gates MC et al. Evidence in practice - a pilot study leveraging companion animal and equine health data from primary care veterinary clinics in 9. New Zealand. Front Vet Sci 2016;3:116. De Briyne N, Atkinson J, Pokludova L, Borriello SP. Antibiotics used most commonly to treat animals in Europe. Vet Rec 2014;175:325. 10. Sparham SJ, Kwong JC, Valcanis M et al. Emergence of multidrug resistance in locally - acquired human infections with Salmonella Typhimurium in Australia owing to a new 11. clade harbouring blaCTX-M-9. Int J Antimicrob Agents 2017;50:101-105. Vandeweerd JM, Vandeweerd S, Gustin C et al. Understanding veterinary practitioners' decision-making process: implications for veterinary medical 12. education. J Vet Med Educ 2012;39:142-151. Mateus AL, Brodbelt DC, Barber N, Stark KD. Qualitative study of factors associated with antimicrobial usage in seven small animal veterinary practices in the UK. Prev 13. Vet Med 2014;117:68-78. Weese JS. Investigation of antimicrobial use and the impact of antimicrobial use guidelines in a small animal veterinary teaching hospital: 1995-2004. Journal of the 14. American Veterinary Medical Association 2006;228:553-558. Committee for Medicinal Products for Veterinary Use. Reflection paper on off-label use of antimicrobials in veterinary medicine in the European Union. European 15. Medicines Agency, http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/20 17/07/WC500232226.pdf, 2017.

12

3 16. Food and Drug Administration. Extralabel use and antimicrobials. 2014. Retrieved 12/10/17. 17. Australian Bureau of Statistics. Regional population growth, Australia 2014-15. http://www.abs.gov.au/ausstats/[email protected]/Previousproducts/3218.0Main Features152014- 15?opendocument&tabname=Summary&prodno=3218.0&issue=2014- 15&num=&view=. 2016. Retrieved 20/11/2017. 18. Humph M. BVA and FVE fight to save vets’ right to dispense medicines. 2011. Retrieved 12/10/17. 19. Brouwers MC, Kho ME, Browman GP et al. AGREE II: advancing guideline development, reporting and evaluation in health care. CMAJ 2010;182:E839-842. 20. Shiffman RN, Dixon J, Brandt C et al. The GuideLine Implementability Appraisal (GLIA): development of an instrument to identify obstacles to guideline implementation. BMC Med Inform Decis Mak 2005;5:23.

12

4 BACK TO TABLE OF Summary CONTENTS

Antimicrobial stewardship measures in veterinary practice have had little development in Australia and few veterinary practices have AMS policies and antimicrobial use guidelines have had minimal development. This thesis has made an important contribution to the literature in documenting areas where antimicrobial use is inappropriate and identifying the challenges in implementing AMS policies and procedures. Importantly, this thesis will lead to practical, and potentially policy change in Australia.

The rate of antimicrobial prescribing in the companion animal community is discussed in chapter 1. Rates of antimicrobial use were lower in cats than in dogs, and both were much lower than the human population in Australia. However, although antimicrobial use was relatively low, inappropriate antimicrobial use was still identified in all sectors investigated in this research (chapters 2-6).

Low antimicrobial dosing was identified as a common reason for inappropriate antimicrobial use, particularly in equine and bovine practice, and antimicrobial drug labelling has been highlighted as one of the potential causes of this. Changes to legislation are needed to ensure that appropriate doses are promoted on the labels, and that with-holding periods for meat and milk are accurate when therapeutic doses are used in food-producing species.

Inappropriate timing, and long durations of antimicrobial therapy, for surgical prophylaxis were also commonly identified and are not a problem restricted to veterinary practice; the medical profession also finds poor compliance with guidelines for surgical prophylaxis. This was a consistent reason for inappropriate use of antimicrobials and highlights the need for AMS programs in veterinary practice.

Recent graduates, and veterinarians from smaller practices, were less compliant with guidelines. For recent graduates, this likely reflects a lack of confidence in decision making, a hypothesis that was supported by further research. Veterinarians from smaller practices may have less access to continuing education and less peer support. Identification of the enablers to, and barriers of, AMS (chapter 7) guided the development of the AMS policy and procedure document (appendix 2) and prompted the investigation of AMS practices in the veterinary diagnostic laboratories. Development of these resources, along with the antimicrobial use guidelines (appendix 1), will lead to practical change within veterinary practices. Future evaluation of these resources and programs will be very exciting!

In conclusion, this thesis has paved the way for the introduction of AMS into veterinary practices. While it is likely that governance will be required for wide- spread implementation of such programs, building an evidence base for the effectiveness of AMS programs is still required and will be a priority of my future work in this area.

1

25 BACK TO TABLE OF Appendix 1 CONTENTS Companion animal poster bacterial enteropathogens TREATMENT SECOND LINE FIRST LINE:______CLINIC POLICY PRESCRIBING GUIDELINES AUSTRALIAN VETERINARY SECOND LINE:______FIRST LINE:______CLINIC POLICY Duration of therapy: one dose only or 2nd dose 6 hours later. CLEAN CONTAMINATED (except dermatitis [treat until cured] and implants [7 days]). Duration of therapy: stop within 24 hours. not tolerate DENTALS WITH EXTRACTIONS: ROUTINE DENTALS CONTAMINATED:______CLEAN CONTAMINATED:______CLEAN:______CLINIC POLICY CONTAMINATED Duration of therapy: no evidence, 24 (PYOMETRA, PROSTATIC ABSCESS, SIGNIFICANT BOWEL LEAKAGE) CONTAMINATED Duration of therapy: stop within 24 hours. CLEAN TIMING until cured. MITIGATING FACTORS: • • • • • • • • • • FIRST LINE Sepsis Specific FIRST LINE IIAIGFACTORS: amoxycillin / clavulanate or 1 MITIGATING FIRST LINE FIRST LINE FIRST LINE SURGERY, NO MITIGATING FACTORS Patients with systemic illness Patients with heart disease Geriatrics Immunosuppressed Surgery involves implant Bacterial dermatitis Endocrine disorder Obese dogs Surgical duration >90 mins Hypotension IV antimicrobials: 30 transient mxcli etmcn+metronidazole + gentamicin + : Amoxycillin Antimicrobials only when signs of sepsis or confirmation of specific :______atra enteropathogens: bacterial ACUTE GASTROENTERITIS SURGERY SURGERY amoxycillin NONE NO ANTIMICROBIALS NONE amoxycillin or 1 amoxycillin or 1 and metronidazole bacteraemia . SURGERY DENTAL SURGERY : use antimicrobial appropriate for infection and treat SC amoxycillin / clavulanate: 2 hours prior to surgery. 4 hours, amoxycillin; every 2 hours. - 60 mins prior to surgery, repeat cefazolin; every cephalosporin IV or IM 30 mins prior to surgery Prophylactic (~20 mins). Recommended for: SURGERY - 48 hours is common in human medicine. st st generation cephalosporin generation cephalosporin metronidazole antimicrobials only in patients that can st generation GROUP 3:______GROUP 2:______GROUP 1:______CLINIC POLICY 3. 2. 1. 3 CATEGORIES TREATMENT Consult with microbiologist to interpreting results (airway contaminants possible). Consider underlying disease process that predisposed to pneumonia. recommended prior to antimicrobial therapy. Tracheal wash for cytology and culture & susceptibility testing is strongly DIAGNOSTICS PNEUMONIA & SEPSIS:______MILD ASPIRATION:______MILD:______CLINIC POLICY Duration of therapy: review after 10 hair follicles and surrounding skin involved: 1 Systemic antimicrobials Surface, superficial, and localised deep pyoderma. TREATMENT Consider underlying disease. Culture and susceptibility testing strongly encouraged when: Cytological evaluation is needed to identify the existence of a bacterial pyoderma. DIAGNOSTICS SECOND LINE:______FIRST LINE:______CLINIC POLICY amoxycillin / Severe bloody diarrhoea with hypovolaemia and sepsis. Severe bloody diarrhoea with hypovolaemia but not septic. Mild bloody diarrhoea, • • • • • • FIRST LINE FIRST LINE FIRST LINE antimicrobials are being considered Recommended in all cases of bacterial pyoderma in which systemic Chronic or recurrent pyoderma New lesions develop during treatment Lack of response to antimicrobial therapy Rods are present on cytology Adhesive tape, direct smear or FNA (pustules or nodules) . clavulanate ACUTE HAEMORRHAGIC ru :fudteayadprnea antimicrobials: and parenteral fluid therapy Group 3: Group 2: fluid therapy and monitor for sepsis Group 1: no antimicrobials Mild aspiration: no treatment or amoxycillin or 1 Mild: doxycycline edn utr n ucpiiiyresults and susceptibility culture pending enrofloxacin and amoxycillin sepsis: & Pneumonia generation cephalosporin for 5 topical shampoo treatment, allow contact with skin – omvlei and systemically well. normovolaemic in cases where large areas of body affected or when - 10 mins PNEUMONIA DIARRHOEA PYODERMA Amoxycillin & gentamicin & metronidazole -14 days. st generation cephalosporins or st SECOND LINE:______FIRST LINE:______CLINIC POLICY resistance. Culture urine before starting treatment. indwelling urinary catheter in dogs or cats. Studies suggest this may promote No evidence to support use of antimicrobials before, during or after removal of an Side effects can occur with long term trimethoprim/ sulphonamide. Urine culture should be performed at 5 Duration of therapy: 2 success and check for recurrence. Urine culture should be performed 7 days after cessation of treatment to confirm Duration of therapy: 5 In OLDER CATS with predisposing diseases only 10-22% will have bacterial cystitis. cystitis. REMEMBER in YOUNG CATS very few with urinary tract signs will have bacterial TREATMENT If complicated, consider underlying disease. Culture and susceptibility testing recommended in all cases. Urinalysis and cytological evaluation of stained and unstained urine sediment. DIAGNOSTICS FOR MORE INFORMATION Interpreting cytology and culture and susceptibility testing difficult. CANINE INFECTIOUS RESPIRATORY DISEASE COMPLEX CHRONIC RHINITIS:______ACUTE RHINITIS:______CLINIC POLICY Duration of therapy: 1 week past resolution of clinical signs. effective than doxycycline or amoxycillin. No evidence that 3 Antimicrobials should be selected based on culture and susceptibility testing. FELINE RHINITIS > 10 days Duration of therapy: 7 Limited benefit of cytology or culture & susceptibility testing. FELINE RHINITIS ≤ 10 days SECOND LINE:______FIRST LINE:______CLINIC POLICY • LOWER URINARY TRACT DISEASE sulphonamide COMPLICATED IN DOGS AND CATS: amoxycillin or trimethoprim/ (pending culture and susceptibility testing) rmtorm/slhnmd r1 or sulphonamide / trimethoprim or amoxycillin CATS: AND DOGS IN UNCOMPLICATED IDIOPATHIC CYSTITIS OF CATS: no antimicrobial therapy mxcli 7 doxycycline or unwell: amoxycillin clinically & pneumonia of evidence No No evidence of pneumonia & clinically well: NONE uouueto uuetbtsseial nel doxycycline unwell: systemically but purulent or Mucopurulent purulent but systemically well: none or Mucopurulent Serious discharge: NONE Collect via UPPER RESPIRATORY DISEASE DOGS & CATS DOGS rd cystocentesis generation cephalosporins or fluoroquinolones are more - - - -10 days. 4 weeks. 7 days. 10 days (pending culture and susceptibility testing) - 7 day intervals. st generation cephalosporin : www.fvas.unimelb.edu.au/ Culture and susceptibility testing recommended when: History, clinical presentation & cytology. DIAGNOSTICS TREATMENT If Systemic antimicrobials only when: SECOND LINE:______FIRST LINE:______CLINIC POLICY Amoxycillin or ampicillin for 5 SECOND LINE:______FIRST LINE:______CLINIC POLICY Less ototoxic agents: fluoroquinolones (marbofloxacin, ciprofloxacin). Ototoxic agents: polymixin B, aminoglycosides. Systemic antimicrobials Duration of therapy: 10 Duration of therapy: 10 pressure. Ear flushing (under GA if necessary): warm sterile saline under controlled TREATMENT anatomical anomaly). If recurrent underlying disease should be investigated (foreign body, atopy, Collect specimens before flushing. Ensure tympanic membrane is intact, ear flushing under GA may be necessary. Culture and susceptibility testing should be performed when: inflammatory cells. Cytological evaluation should always be performed to identify pathogens and DIAGNOSTICS Non middle or inner ear is involved. Base therapy on culture and susceptibility. doesn’t respond consider underlying disease. -ototoxic agents: chlorhexidine, • • • Rods only: oc nyOR cocci & rods: only Cocci FIRST LINE FIRST LINE - Intact tympanic membrane: ear flushing, topical therapy with - - Intact tympanic membrane: ear flushing, topical therapy with - fucidic Chronic otitis Lack of response to antimicrobial therapy Rods are present on cytology Perforated tympanic membrane: ear flushing and non Perforated tympanic membrane: ear flushing and non polymixin laes vi oia antimicrobials topical avoid cleaners, cleaners, avoid topical antimicrobials AND TRAUMATIC WOUNDS Lack of response to antimicrobial therapy Immunosuppressed patient Potential joint involvement Diffuse tissue involvement Systemically unwell acid and CELLULITIS, ABSCESS ,gnaii or gentamicin B, Draining -14 days. -14 days. – OTITIS EXTERNA framycetin often ineffective and usually only indicated when -10 days. & flushing alone Tris obnto rgentamicin or combination marbofloxacin - EDTA. vetantibiotics -ototoxic -ototoxic

12

6 BACK TO TABLE OF Large animal flip book CONTENTS

AUSTRALIAN VETERINARY CATTLE PRESCRIBING GUIDELINES DOSE RATES

ANTIMICROBIAL AGENT RECOMMENDED ROUTE INTER-DOSING WITHHOLDING DOSE INTERVAL PERIOD (days)

Procaine penicillin* 22,000 IU/kg IM 12 - 24 hours Not established, test

Oxytetracycline 10 mg/kg IV or IM 12 - 24 hours Milk: 5 Meat: 14

Oxytetracycline long 20 mg/kg IM 72 hours Milk: 7 Meat: 28 acting Amoxycillin / clavulanate 10 mg/kg PO 12 hours Meat: 4 (pre-ruminant calves)

Trimethoprim / 24 mg/kg IM 12 - 24 hours Milk: 3 Meat: 28 sulphonamide

Tulathromycin 2.5 mg/kg SC Once Meat: 35 (beef and dairy heifers)

Florfenicol 40 mg/kg SC Once Meat: 55 (not in dairy cattle) 20 mg/kg IM 48 hours Meat: 36 *Many of the recommendation in this guide represent off-label use of antimicrobials. Compliance with the legal requirements of your jurisdiction is your responsibility.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V CATTLE P G SURGICAL PROPHYLAXIS

SURGICAL ANTIMICROBIAL DURATION OF CONTAMINATION RECOMMENDATION THERAPY LEVEL

CLEAN, NO MITIGATING NONE N/A FACTORS

CLEAN, MITIGATING FACTORS Oxytetracycline Stop within 24 hours

CLEAN CONTAMINATED Oxytetracycline 24-48 hours

Choose antimicrobial appropriate for CONTAMINATED Treat till cured infection MITIGATING FACTORS TIMING: Tissue levels are required at the time of incision • Surgical duration >90 mins to confer protection from surgical site infection. • Rumenotomy IV antimicrobials: 30-60 minutes prior to surgery • Unsanitary conditions IM oxytetracycline: 8 hours prior to surgery • Periparturient IM penicillin: 2 hours prior to surgery

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

7 A V CATTLE P G FOOT DISEASE

FOOTROT FOOT ABSCESS

DIAGNOSTICS DIAGNOSTICS Diagnosis can be made from clinical signs alone. The foot Diagnosis can be made from examination of the foot. must be lifted for examination in all cases. Ensure no foreign body is present in the interdigital space. TREATMENT Antimicrobials are not indicated. Establishing drainage is the critical TREATMENT factor. Topical therapy with antibacterial disinfectant. Procaine penicillin is highly effective. Florfenicol is a suitable alternative in beef cattle.

DURATION OF THERAPY DIGITAL DERMATITIS A single dose of florfenicol or 3 days of procaine penicillin is generally sufficient. Treat until lesions have resolved. “Hairy Heelwart” DIAGNOSTICS Diagnosis can be made from examination of the foot.

TREATMENT Topical therapy with tetracycline is most effective. Bandaging maintains tetracycline contact with lesions.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V CATTLE P G RESPIRATORY

PNEUMONIA CALF DIPTHERIA

DIAGNOSTICS DIAGNOSTICS Most common pathogens are Mannheimia haemolytica, Diagnosis usually based on clinical signs. Pasteurella multocida, Histophilus somni and Mycoplasma spp, Consider underlying disease (persistently infected with BVD) or often in conjunction with viral pathogens. foreign body. Although diagnostics are rarely pursued, they should be considered for valuable animals or in outbreaks. TREATMENT Culture and susceptibility testing can be performed from Procaine penicillin is preferred. Oxytetracycline is suitable transtracheal wash, bronchoalveolar lavage or post-mortem alternative. specimens. Mannheimia can be associated with Severe cases may require tracheotomy. pleuropneumonia, which carries a very poor prognosis. DURATION OF THERAPY TREATMENT 5 days of procaine penicillin or 2 doses of long acting Oxytetracycline most appropriate. Ceftiofur has limited activity oxytetracycline (3 days apart) is generally sufficient. against Mycoplasma spp.

DURATION OF THERAPY Dependent on severity. 2-3 days may be adequate in mild cases. Treat until disease resolved, which may take > 1 week in severe cases.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

8 A V CATTLE P G GASTROINTESTINAL

CALF DIARRHOEA ENTERITIS IN ADULTS PERITONITIS DIAGNOSTICS DIAGNOSTICS DIAGNOSTICS Rapid (patient side) diagnostics, Faeces should be submitted for culture and Abdominocentesis is recommended for performed on faeces, are available and susceptibility testing if salmonellosis is cytological evaluation at a minimum and should be utilised to confirm bacterial suspected. preferably also for culture and origin as most are not. E. coli (< 3 days of susceptibility testing. age) and Salmonella are possible bacterial TREATMENT Consider origin of bacterial contamination causes. Antimicrobial therapy is not indicated for as this affects prognosis. enteritis in adult cattle that are systemically TREATMENT well. TREATMENT Antimicrobial therapy is not indicated for Systemic antimicrobials are indicated when: Broad-spectrum coverage is required as a diarrhoea caused by viruses or • Invasive salmonellosis is suspected mixed population of bacteria are usually crytosporidia. Systemic antimicrobials are • Signs of sepsis present, including anaerobes. indicated when: Trimethoprim / sulphonamide or Oxytetracycline is preferred. Trimethoprim • Documented bacterial aetiology oxytetracycline are suitable choices. / sulphonamide is a suitable alternative. • Sepsis Both should be used twice daily. • High-risk of sepsis DURATION OF THERAPY Trimethoprim / sulphonamide or 5 days is generally considered adequate. DURATION OF THERAPY oxytetracycline are suitable choices. Dependent on severity. Mild cases (post- surgery) may respond in 5 days. GI DURATION OF THERAPY contamination (i.e. following rupture of an 5 days is generally considered adequate. abomasal ulcer) may require 2-3 weeks of therapy.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V CATTLE P G MASTITIS

GRAM NEGATIVE, SEVERE GRAM POSITIVE

DIAGNOSTICS DIAGNOSTICS Diagnosis is generally made from clinical signs alone. Milk Milk samples should be obtained for somatic cell count and for culture samples should be obtained for culture and susceptibility and susceptibility testing, especially in an outbreak. Samples can be testing. frozen, for later submission, if empirical treatment fails. Training of farmers on aseptic milk collection techniques is critical. TREATMENT Antimicrobial therapy should be initiated immediately TREATMENT following sample collection as the disease is rapidly Intramammary antimicrobials are preferred as they exert less pressure progressing, and often fatal. on resistance development at a farm level. Antimicrobial selection Oxytetracycline should be administered intravenously as should be guided by culture and susceptibility results. Preparations perfusion of the muscles is often poor so drug absorption is containing cloxacillin or amoxycillin are generally effective against reduced. Intramammary therapy has poor penetration. Streptococcus spp. (most frequently cultured organisms). Supportive therapy is strongly recommended (fluid therapy Staphylococcus aureus is associated with biofilm formation, which and non-steroidal anti-inflammatory drugs). worsens the prognosis. Treatment during lactation may not be successful. DURATION OF THERAPY If indicated, preferred systemic antimicrobials are penethamate 5-7 days generally required. hydrochloride and trimethoprim / sulphonamide.

DURATION OF THERAPY Treat until clinical signs resolve and milk somatic cell count is normal. 2-3 days may be sufficient for mild cases.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

9 A V CATTLE P G NEUROLOGICAL

LISTERIA THROMBOEMBOLIC OTITIS MEDIA MENINGOENCEPHALITIS

DIAGNOSTICS DIAGNOSTICS DIAGNOSTICS Diagnosis is generally made from Pneumonia is generally present concurrently, or Frequently secondary to pneumonia in clinical signs. CSF can be collected for in other in-contact animals, helping to calves kept in poorly ventilated areas. cytological evaluation, culture is rarely differentiate this disease from listeriosis. Diagnosis can be made from clinical successful. signs alone. TREATMENT TREATMENT Intravenous oxytetracycline is strongly TREATMENT Intravenous oxytetracycline or recommended. Twice daily dosing is advised. Oxytetracycline is preferred. crystalline penicillin is strongly Tulathromycin is a suitable alternative. recommended. Twice daily dosing is DURATION OF THERAPY advised. 5-7 days is generally recommended. DURATION OF THERAPY 3-5 days of oxytetracycline is generally DURATION OF THERAPY required. A single dose of tulathromycin is 5-7 days is generally recommended. sufficient.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V CATTLE P G MISCELLANEOUS

METRITIS NEONATAL PINKEYE SEPTICAEMIA DIAGNOSTICS DIAGNOSTICS DIAGNOSTICS Diagnosis is generally made from clinical Diagnosis is generally made from clinical Diagnosis is generally made from clinical signs alone. signs. signs. Consider bacterial aetiology (enteritis, TREATMENT omphalophlebitis/naval ill) or failure of TREATMENT Systemic antimicrobials should only be passive transfer. Topically therapy with cloxacillin is used when severe systemic illness is generally effective. Use of present. Antimicrobial therapy is not TREATMENT ophthalmological formulations is preferred indicated in cattle that are clinically well. Oxytetracycline can be used but care should as the duration of action is longer. Oxytetracycline is preferred. be taken with hypovolaemic animals as Subpalpebral administration of penicillin is Supportive therapy may be required (fluid renal toxicity can occur. Trimethoprim / useful in severe cases. therapy and non-steroidal anti- sulphonamide is a suitable alternative. Covering the eye with a patch aids in inflammatory drugs). recovery and reduces transmission of DURATION OF THERAPY disease. DURATION OF THERAPY 5-7 days may be sufficient in uncomplicated 3 days is generally sufficient but longer disease. Longer durations are necessary DURATION OF THERAPY may be necessary in severe cases. when omphalophlebitis or septic arthritis One application of cloxacillin ointment may develop. Up to 2 weeks may be necessary. be sufficient. Severe cases may need treatment every 48 hours (1-2 additional applications).

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

30 A V HORSES P G DOSE RATES

ANTIMICROBIAL RECOMMENDED ROUTE INTER-DOSING AGENT DOSE INTERVAL

Procaine penicillin* 22,000 IU/kg IM 12 hours

Gentamicin* 7.7-9.7 mg/kg IV or IM 24 hours

Trimethoprim / 30 mg/kg PO or IV 12 hours sulphonamide

Doxycycline* 10 mg/kg PO 12 hours

Oxytetracycline* 6.6 mg/kg Slow IV 12 hours

Metronidazole* 20mg/kg PO 12 hours

*Many of the recommendation in this guide represent off-label use of antimicrobials. Compliance with the legal requirements of your jurisdiction is your responsibility.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V HORSES P G SURGICAL PROPHYLAXIS

SURGICAL ANTIMICROBIAL DURATION OF CONTAMINATION RECOMMENDATION THERAPY LEVEL CLEAN, NO MITIGATING NONE N/A FACTORS

CLEAN, MITIGATING FACTORS Penicillin & Gentamicin Stop within 24 hours

CLEAN CONTAMINATED Penicillin & Gentamicin 24-48 hours

CONTAMINATED Choose antimicrobial appropriate for Treat until cured infection

MITIGATING FACTORS TIMING: Tissue levels are required at the time of incision • Surgical duration >90 mins to confer protection from surgical site infection. • Surgery involving an implant IV antimicrobials: <60 minutes prior to surgery • Surgical site infection would be a IM procaine penicillin: 3.5 hours prior to surgery major threat to the patient (i.e. central nervous system surgery) FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

31 A V HORSES P G SKIN/FEET

WOUNDS FOOT ABSCESS CELLULITIS

NO SYNOVIAL STRUCTURES No antimicrobial therapy indicated. PRIMARY no obvious underlying cause. INVOLVED: no antimicrobials therapy Often more severe than secondary cases. indicated, even if contamination of the Curette to establish drainage. If recurrent SECONDARY: an underlying cause can be wound is present. consider underlying disease. Radiographs identified (surgery, joint injection, wound, Systemic antimicrobials only when: should be taken to investigate for pedal blunt trauma). • Systemically unwell osteitis & ACTH measured to investigate for • Potential synovial involvement (see equine Cushing’s disease (PPID). DIAGNOSTICS below) Systemic antimicrobials only when: Fine-needle aspirate should be collected for • Immunosuppressed patient • Immunosuppressed patient culture and susceptibility testing. Care if • If severe cellulitis is present needed for cellulitis occurring over synovial SYNOVIAL STRUCTURE INVOLVED: structures. Lavage is almost always required for successful outcome. Systemic Ensure horses are vaccinated for tetanus. TREATMENT antimicrobials always indicated. Therapy IVRP: gentamicin 1/3 systemic dose should be based of culture and Systemic antimicrobials: Penicillin & susceptibility testing. Empirical therapy gentamicin (adjust dose if IVRP performed) with penicillin and gentamicin should be or oxytetracycline. initiated pending culture results. Topical therapy: Cold water hosing and pressure bandage. Analgesia especially if non-weight bearing as risk laminitis in contralateral limb.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V HORSES P G RESPIRATORY

STRANGLES SINUSITIS PNEUMONIA

DIAGNOSTICS DIAGNOSTICS DIAGNOSTICS Notifiable disease, samples should be A sample of fluid from the sinus should be Transtracheal wash, or endoscopic submitted for serology, culture or PCR to obtained to confirm the diagnosis. Culture tracheal wash with a triple guarded confirm diagnosis. is not usually required. catheter, should be performed for Consider underlying disease (dental or cytological evaluation. Culture and TREATMENT equine Cushing’s) especially if recurs. susceptibility testing should be performed No antimicrobial recommended. Most in all cases. Culture of bronchoalveolar cases resolve quickly once drainage has TREATMENT lavage specimens is never appropriate as been established. A small percentage Sinus lavage alone may be sufficient and is these samples are contaminated by the continue to shed (carriers). almost always required for successful upper airway. Systemic antimicrobials only when: outcome (minimally invasive technique in • Respiratory compromise the field can be used). TREATMENT • Metastatic disease (Bastard strangles) Systemic antimicrobials when: Should be based on culture and In these cases, penicillin is first line • Recurrent disease susceptibility results. Empirical therapy with therapy. • Systemically unwell penicillin & gentamicin should be initiated In these cases, penicillin or trimethoprim / pending results. Metronidazole should be sulphonamide is first line therapy. added if anaerobes are suspected (foul smell to tracheal fluid).

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

32 A V HORSES P G FOALS

PNEUMONIA SEPSIS SEPTIC ARTHRITIS

DIAGNOSTICS DIAGNOSTICS DIAGNOSTICS Streptococcus zooepidemicus and Sepsis score can be used to assess risk Arthrocentesis should be performed to Rhodococcus equi are equally common. (see website). Blood for culture and obtain fluid for cytological evaluation and for Transtracheal wash is required for susceptibility should be collected but culture and susceptibility testing in all cytological examination and culture and false negatives are common. cases. Radiographs should be taken to susceptibility testing in all cases. investigate bone involvement. TREATMENT TREATMENT Based on culture and susceptibility TREATMENT Based on culture and susceptibility results. results if possible. Empiric therapy can be Based on culture and susceptibility results. Empiric therapy can be initiated while initiated while results pending. Empiric therapy can be initiated while results pending. If S. zooepidemicus is Penicillin & gentamicin is recommended. results pending. Penicillin & gentamicin is suspected penicillin is appropriate. If R. Care with gentamicin if renal function is recommended. Oxytetracycline is an equi is suspected clarithromycin and compromised. Intravenous trimethoprim / alternative, especially if osteomyelitis is rifampin is recommended. sulphonamide is alternate. diagnosed.

DURATION OF THERAPY DURATION OF THERAPY DURATION OF THERAPY Varies by pathogen; 1 week generally 2 weeks is generally considered to be Treat for 1 week past resolution of clinical adequate for S. zooepidemicus, 4-6 week adequate, unless focal infection develops signs, longer if osteomyelitis is present. generally recommended for R. equi. (i.e. septic arthritis).

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V HORSES P G FOALS

PATENT URACHUS OMPHALOPHLEBITIS HIGH-RISK FOALS (NAVEL ILL) DIAGNOSTICS DIAGNOSTICS Premature foal and those with neonatal Ultrasound evaluation should be Ultrasound evaluation should be performed encephalopathy (’Dummy Foal Syndrome’) performed to rule out omphalophlebitis. to define the infected structure and to allow are at increased risk of sepsis. Failure of If no enlargement of the umbilical for monitoring with treatment. passive transfer should be addressed with remnants is identified antimicrobial plasma transfusion. There is no evidence for therapy is not indicated. TREATMENT any benefit from prophylactic antimicrobials in Penicillin & gentamicin is most effective but place of plasma transfusion. TREATMENT often not tolerated well. Trimethoprim / No antimicrobial therapy indicated. sulphonamide or doxycycline are suitable DIAGNOSTICS Frequent topical antibacterial therapy alternatives that can be given orally. Serial haematologic evaluation and sepsis with chlorhexidine is recommended until score may guide necessity for antimicrobial patency resolves. DURATION OF THERAPY therapy. Serial ultrasonographic examination should be performed and therapy continued until 1 TREATMENT week after resolution of disease. Prophylactic therapy is warranted when leukopaenia is present or sepsis score is high. Penicillin & gentamicin is most appropriate but care should be taken in foals with impaired renal function. Trimethoprim / sulphonamide IV is an alternative.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

33 A V HORSES P G GASTROINTESTINAL

LAWSONIA DIARRHOEA PERITONITIS (PROLIFERATIVE ENTEROPATHY) ACUTE DIARRHOEA DIAGNOSTICS DIAGNOSTICS DIAGNOSTICS Abdominocentesis should be performed to Diagnosis can be made via serology Culture should be performed for collect fluid for cytological evaluation and (ELISA) or by faecal PCR. Salmonella. Diagnosis of clostridial culture and susceptibility testing. disease requires toxin test. Differentiation between primary and TREATMENT secondary origins is critical as secondary Mild to moderate disease: doxycycline PO TREATMENT peritonitis is typically due to leakage from Severe disease: oxytetracycline IV Antimicrobial therapy rarely indicated. the gastrointestinal or reproductive tracts Only if: and surgery should be considered. DURATION OF THERAPY • Confirmed clostridial cause Mild to moderate disease: generally 3 • Severe leukopaenia and neutropaenia TREATMENT weeks is recommended If clostridial: metronidazole Systemic antimicrobial therapy should be Severe disease: 3-4 weeks If leukopaenic: penicillin & gentamicin instituted immediately following sample collection. Penicillin & gentamicin & DURATION OF THERAPY metronidazole are appropriate. Clostridial: until diarrhoea resolves Leukopaenic: until leukopaenia resolves DURATION OF THERAPY Serial abdominocentesis should guide CHRONIC DIARRHOEA therapy. Treat for 1-2 weeks past resolution Antimicrobial therapy rarely indicated. of disease

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

A V HORSES P G REPRODUCTION

RETAINED PLACENTA PLACENTITIS ENDOMETRITIS

DIAGNOSTICS DIAGNOSTICS DIAGNOSTICS Diagnosis can be made on clinical signs Ultrasonographic examination of the Cytological evaluation and culture and alone. placenta is necessary. Samples should be susceptibility testing is required for diagnosis. collected for culture and susceptibility Consider underlying disease. TREATMENT testing if the cervix is open. There is no Large volume uterine lavage is critical for evidence for prophylactic or pulse therapy TREATMENT stimulating placental detachment and for placentitis. There is no evidence for routine treatment of removing endotoxins thereby preventing mares post-service. Therapy should be absorption. Systemic antimicrobials are TREATMENT guided by culture and susceptibility results. always required. Penicillin, gentamicin Trimethoprim / sulphonamide is preferable Intrauterine penicillin and aminoglycoside and metronidazole should be and gentamicin may not cross the placenta. appears effective in most cases. administered. NSAIDs are also critical. DURATION OF THERAPY DURATION OF THERAPY 1 week past resolution of ultrasonographic 1 week past resolution of clinical and clinical disease or until foaling. disease. Generally requires therapy until foaling.

FOR MORE INFORMATION: www.fvas.unimelb.edu.au/vetantibiotics

34 BACK TO TABLE OF Appendix 2 CONTENTS

Antimicrobial stewardship policy

Veterinary Antimicrobial Stewardship Policy

POLICY SUMMARY: This policy provides the framework for how the antimicrobial stewardship program operates in veterinary practices. This program aims to promote the optimal use of antimicrobial agents by veterinarians in order to maximise effectiveness and minimise the potential for drug resistance.

1. PURPOSE AND SCOPE Antimicrobial resistance is a global health emergency. The purpose of the antimicrobial stewardship policy is to improve veterinary antimicrobial stewardship and encourage appropriate antimicrobial use in line with the recommendations set out in the Implementation Plan for the National Antimicrobial Resistance Strategy (2015-2019).

2. POLICY 2.1 Veterinarians and human healthcare professionals share the responsibility to prevent the development and transmission of multi-drug resistant organisms by ensuring that effective education and training in appropriate antimicrobial use is undertaken (Objective 1). 2.2 Veterinarians will implement antimicrobial stewardship practices (Objective 2). 2.3 Veterinarians will have access to current, evidence based antimicrobial use guidelines, and/or locally endorsed guidelines (Objective 2.1). 2.4 Veterinarians will promote the awareness and understanding of antimicrobial resistance in animals (Objective 1.2). 2.5 Veterinarians will institute infection prevention and biosecurity measures in their practices to prevent infections and the spread of antimicrobial resistance (Objective 4.5).

2.6 Veterinarians will provide science based advice to their clients on infection prevention and biosecurity measures to prevent infections and the spread of antimicrobial resistance (Objective 4.5).

3. ASSOCIATED PROCEDURES 3.1 Veterinary antimicrobial stewardship program 3.2 AVA Personal Biosecurity Guidelines

4. REFERENCES 4.1 ACSQHC Antimicrobial Stewardship in Australian Hospitals 4.2 Antimicrobial stewardship in outpatient settings: A systematic review

1

35 BACK TO TABLE OF Antimicrobial stewardship procedure template CONTENTS [ENTER CLINIC NAME] Veterinary Antimicrobial Stewardship Program

[This document can be adapted to suit your clinic. We recommend maintaining the essence of the document as this represents the best adaption of the evidence to date.]

PURPOSE:

To comply with evidence-based guidelines or best practices regarding antimicrobial prescribing and promote rational and appropriate antimicrobial therapy while improving clinical outcomes and minimising unintentional side-effects of antimicrobial use, including emergence of resistant microorganisms. The program is based on the ACSQHC Antimicrobial Stewardship in Australian Hospitals in medical practice and a 2015 systematic review of antimicrobial stewardship in outpatient settings and has been adapted by the University of Melbourne to reflect practicalities in veterinary practice in Australia. The outcomes (www.fvas.unimelb.edu.au/vetantibiotics) and impact of the program should be tracked by clinic management to ensure implementation of the program. [This document outlines the structure recommended for antimicrobial stewardship in veterinary practices. You may choose a program level or individual aspects of the program from different levels. Detailed explanations of the program components can be found on subsequent pages of this document. Use the checklist on the next page to select the components to be implemented in your practice. We strongly recommend choosing at least 5 components. We have outlined three possible programs in Table1. The Bronze program reflects a level of specialist knowledge and access to essential resources. The Silver program includes appointment of an antimicrobial stewardship champion and active use of several interventions. The Gold program includes all of the features of Bronze and Silver programs but includes additional monitoring, review and feedback procedures. ]

13

6 Table 1. Suggested outline of 3 levels of antimicrobial stewardship program Program Resources Education Active interventions tools

Program level Bronze Antimicrobial use Participation in continuing Understand the principles guidelines. education on of judicious antimicrobial use by at antimicrobial use. least one veterinarian on Knowledge of, and access staff annually. to, antimicrobial use guidelines Silver Antimicrobial use Education of staff at Follow antimicrobial use guidelines. induction. guidelines. Antimicrobial Participation in continuing Understand and follow stewardship education on principles of judicious champion antimicrobial use by one antimicrobial use. appointed. or more veterinarians on Institue a traffic-light staff annually. system for antimicrobial Educate clients on judicious use. antimicrobial use in Institute diagnostic testing animals. guidelines. Gold All of the All of the recommendations All of the recommendations for silver and: recommendations for for silver and: Participation in continuing silver and: Local guidelines for education on Institute audit and antimicrobial use. antimicrobial use by half feedback procedures. of all veterinarians on Restrict access to staff annually. antimicrobials with Education of staff at high-importance rating. induction and, at least, Utilise delayed annually thereafter. prescribing. All clinic meetings to Monitor antimicrobial include antimicrobial resistance and stewardship as an agenda antimicrobial use. item.

13

7 ELEMENTS OF VETERINARY ANTIMICROBIAL STEWARDSHIP PROGRAM

INTERVENTIONS [select at least five interventions]

Follow antimicrobial use guidelines Implement clinic policy on antimicrobial use Appoint antimicrobial stewardship champion Use a traffic-light system for antimicrobial use Use diagnostic testing guidelines Restrict access to antimicrobials with high-importance rating Use delayed prescribing Monitor antimicrobial use Monitor antimicrobial resistance Other [specify]

EDUCATION [select at least one education strategy]

Education of staff about antimicrobial use and clinic policies at induction and, at least, annually thereafter Participation in continuing education on antimicrobial use by one or more staff veterinarians annually Educate clients on judicious antimicrobial use in animals All clinic meetings to include antimicrobial stewardship on the agenda Other [specify]

RESOURCES [select resources that will be available for staff and where they can be accessed]

Antimicrobial use guidelines Antimicrobial stewardship champion Clinic policy for antimicrobial use Other [specify]

**BIOSECURITY POLICIES SHOULD ALSO BE IMPLEMENTED

A biosecurity policy and procedures document should be produced and accompany this document. The AVA produces guidelines.

13

8 RESOURCES:

Antimicrobial Guidelines & judicious antimicrobial use

Bronze: A copy of the antimicrobial guidelines will be available for all staff. The poster for companion animal practices and booklets or posters for large animal practices (resources available at www.fvas.unimelb.edu.au/vetantibiotics) will be available. Judicious use of antimicrobials will be promoted at staff meetings.

Silver: As above AND the antimicrobial stewardship champion, in consultation with all participating veterinarians, will utilise the guidelines to develop recommendations that shall be populated in the poster for companion animal practices and in booklets or posters for large animal practices (resources available at www.fvas.unimelb.edu.au/vetantibiotics). Judicious use of antimicrobials will be promoted at staff meetings.

Gold: As above AND the antimicrobial stewardship champion, in consultation with all participating veterinarians, will utilise evidence-based practice protocols and guidelines to develop clinic protocols for antimicrobial use. Recommendations shall be populated in the poster for companion animal practices and in booklets or posters for large animal practices (resources available at www.fvas.unimelb.edu.au/vetantibiotics). The antimicrobial stewardship champion should access antibiograms, if these are available, to guide updates to practice protocols. Judicious use of antimicrobials will be promoted at staff meetings.

Antimicrobial stewardship champion

Silver & Gold: The antimicrobial stewardship champion will be a practice owner, or a senior veterinarian with the support of the practice owners. The antimicrobial stewardship champion should be trained in antimicrobial stewardship.

Educational Activities

Bronze: At least one veterinarian on staff will participate in continuing education about judicious use of antimicrobials annually.

Silver: One or more veterinarians on staff will participate in continuing education about judicious use of antimicrobials annually. Education of staff regarding evidence-based guidelines or best practices including antimicrobial management should occur at induction.

Gold:

13

9 Half of all veterinarians on staff will participate in continuing education about judicious use of antimicrobials annually. Education of staff regarding evidence- based guidelines or best practices including antimicrobial management should occur at induction and, at least, annually thereafter. All clinic meetings should include antimicrobial stewardship as an agenda item.

Client education

Silver & gold: The practice manager or antimicrobial stewardship champion will display educational posters in reception areas to educate clients on responsible use of antimicrobials. In addition, literature such as flyers will be available to assist veterinarians to explain antimicrobial resistance and the reasons when antimicrobials may not be prescribed. All veterinarians will commit to educating clients about responsible and prudent use of antimicrobials for their animals.

Traffic-light system for antimicrobials

Silver & gold: The antimicrobial stewardship champion will coordinate colour coding of antimicrobials held in the pharmacy, either by application of a coloured sticker or by colour coding of the shelves where antimicrobials are stored. Recommended classification of antimicrobials can be found in table 1.

The antimicrobial stewardship champion will add additional antimicrobials to the list as they become available in the veterinary clinic.

1

40 Table 1. Traffic light system for antimicrobials used in veterinary medicine (based on Australian Strategic Technical Advisory Group on AMR antimicrobial rating system) GREEN: FIRST CHOICE ORANGE: SECOND LINE RED: ONLY AFTER C & S* (LOW IMPORTANCE RATING) (MEDIUM IMPORTANCE RATING) (HIGH IMPORTANCE RATING) Penicillins Penicillins Penicillins Benzyl penicillin Amoxycillin / clavulanate Piperacillin / tazobactam Procaine penicillin Cloxaxillin Ticarcillin / clavulanate Benzathine penicillin Dicloxacillin Amoxycillin Flucloxacillin Ampicillin

Tetracyclines 1st Generation Cephalosporins 3rd Generation Cephalosporins Tetracycline Cephalexin Cefovecin Doxycycline Cephalothin Ceftiofur Minocycline Cefazolin Ceftriaxone 2nd Generation Cephalosporins Cefotaxime Cefaclor 4th Generation Cephalosporins Cefoxitin Ceftazidime Cefapime Ceftaroline

Aminoglycosides Aminoglycosides Aminoglycosides Neomycin Gentamicin Amikacin Streptomycin Tobramycin Spectinomycin

Sulphonamides Lincosamides Fluoroquinolones Sulfadiazine Clindamycin Enrofloxacin Silver sulfadiazine Lincomycin Ciprofloxacin Sulfacetamide Marbofloxacin Trimethoprim / Pradofloxacin sulfamethoxazole Ofloxacin Sulfadoxine / pyrimethamine

Macrolides Nitroimidazoles Glycopeptides Azithromycin Metronidazole Vancomycin Clarithromycin Teicoplanin Erythromycin

Polypeptides Pseudomonic acids Carbapenems Bacitracin Mupirocin Imipenem Meropenem

Amphenicols Monobactams Chloramphenicol Aztreonam Antimycobacterials Isoniazid Rifamycins Rifampicin Polymixins Polymixin B Colistin * Or if compelling reason to use pending C & S

Diagnostic testing

Silver & gold:

1

41 All veterinarians will use diagnostic testing when indicated in the guidelines for discernment between diseases with a bacterial origin and those without a bacterial origin. Cases that do not have evidence of a bacterial cause will not be treated with antibiotics unless substantial mitigating factors exist; these will then be documented in the clinical record.

Culture and susceptibility testing will be performed prior to prescribing antimicrobials for: • All cases of suspected urinary tract infection (along with cytological examination) • Any case that does not respond to empirical therapy • Any case where an antimicrobial with a high importance rating is used empirically • Any case where infection is life-threatening (e.g. sepsis, septic arthritis, meningitis) • [add others as appropriate]

Cytology will be performed prior to prescribing antimicrobials for:

• All dermatological & otitis cases • All cases of suspected urinary tract infection • All cases of suspected synovial sepsis • All cases of suspected pleuritis, peritonitis or pneumonia • [add others as appropriate]

Patient side tests will be performed prior to prescribing antimicrobials for:

• All cases of neonatal calf diarrhoea • [add others as appropriate]

Audit & feedback

Gold: The antimicrobial stewardship champion shall review antimicrobial prescribing for recent graduates and new team members using the MIND-ME principle:

M Microbiology should guide therapy where possible

I Indications should be evidence-based and include a bacterial cause in most instances

N Narrowest spectrum possible

D Dosage appropriate to species, site and type of infection (label not always accurate)

M Minimise duration of therapy

E Ensure monotherapy wherever possible (one drug rather than combinations of drugs)

1

42 All veterinarians shall review culture & susceptibility reports and consider potential adjustments to antimicrobial regimens. De-escalation should be considered, which involves switching from drugs with a high importance rating to those with a low or medium importance rating if susceptibility testing indicates their effectiveness, or shifting from broad spectrum to narrow spectrum therapy once a causative agent has been identified. Escalation of therapy may also be required, if empirical antimicrobial therapy is not effective against the causative agent identified on culture. Consultation with a microbiologist may be required and should be considered in any case where the organism was unexpected or has an unusual susceptibility pattern.

Delayed prescribing

Gold: All veterinarians will use delayed prescribing for conditions where there is no evidence that antimicrobials are immediately necessary and other therapies such as topical therapy or anti-inflammatories have been prescribed. Should effective treatment fail with these, antimicrobial therapy would then be indicated. Repeat consultation may be an alternative in some circumstances, but where client finances do not allow for repeat consultation, delayed prescribing is an acceptable alternative. Delayed prescribing typically gives patients a 2-day delay on a prescription, so if symptoms persist or worsen antimicrobials can be obtained.

Antimicrobial class restriction

Gold: A list of restricted antimicrobials shall be made available to all veterinarians with permission for use required from the antimicrobial stewardship champion or [insert names of others as required] before their use.

Alternatively a primary list of restricted antimicrobials shall be made available to all veterinarians with permission for use required from the antimicrobial stewardship champion or [insert names of others as required] before their use, and a secondary list of restricted antimicrobials shall be made available with use only allowed if samples are submitted for culture and susceptibility testing prior to initiating therapy. Recommended classification of antimicrobials for restriction can be found in table 2.

Utilisation of the following antimicrobials will be reviewed annually and presented to all veterinarians in a clinic meeting.

1

43 Table 2. Restricted antimicrobials Restricted Antimicrobials [CAN BE ADAPTED TO INDIVIDUAL CLINICS] Permision required Use allowed if samples for C&S collected prior to starting therapy Piperacillin/tazobactam (Piptaz®) Amikacin Ceftaroline (Teflaro®) Ticarcillin/clavulanate (Timentin®) Colistimethate IV (Colistin®) Tigecycline (Tygacil®) Daptomycin (Cubicin®) Ceftriaxone Cefotaxime, Ceftazidime, Cefapime, Enrofloxacin, Moxifloxacin, Pradofloxacin, Cefotaxime, Ceftazidime, Ceftriaxone Ciprofloxacin Linezolid (Zyvox®) / Tedizolid Rifampicin (Sivextro®) Meropenem (Merrem®) / Imipenem Cefovecin (Convenia®) (Primaxin®) / Doripenem (Doribax®), Ertapenem (Invanz®) Aztreonam Polymixin B (IV) Vancomycin Ceftiofur (Accent®, Excenel®, Excede®) Tiecoplanin Norfloxacin, Ofloxacin, Levofloxacin Isoniazid *Proposed list of restricted antimicrobials.

Surveillance

Gold: Antimicrobial resistance All veterinarians shall report antimicrobial resistance to the antimicrobial stewardship champion. The antimicrobial stewardship champion will record all occurences of resistance, and collate susceptibility patterns to ensure early detection of emergence of drug resistant organisms. Any resistant microbial pathogens that appear to have similar susceptibility profiles may represent nosocomial transmission and warrant prompt investigation and documentation.

Antimicrobial use The antimicrobial stewardship champion will record the case details, indication for use, dose, duration of therapy and outcome of all cases treated with restricted antimicrobials.

1

44

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: Hardefeldt, Laura Yvonne

Title: Antimicrobial stewardship in Australian veterinary practices

Date: 2017

Persistent Link: http://hdl.handle.net/11343/198446

File Description: Complete thesis

Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.