BONE AND JOINT INFECTIONS : from microbiology to diagnostics and treatment. [2 ed.] 9781119720652, 1119720656

Citation preview

Bone and Joint Infections

This book is dedicated to my wife Annelies

Bone and Joint Infections From Microbiology to Diagnostics and Treatment Editor Werner Zimmerli

Second Edition

This edition first published 2021 © 2021 John Wiley & Sons Ltd Edition History First edition © 2015 by John Wiley & Sons, Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Werner Zimmerli to be identified as the author of the editorial material in this work has been asserted in accordance with law. Registered Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Office 9600 Garsington Road, Oxford, OX4 2DQ, UK The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Boschstr. 12, 69469 Weinheim, Germany 1 Fusionopolis Walk, #07‐01 Solaris South Tower, Singapore 138628 For details of our global editorial offices, customer services, and more information about Wiley products visit us at www. wiley.com. Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging‐in‐Publication Data Name: Zimmerli, Werner, 1948– editor. Title: Bone and joint infections : from microbiology to diagnostics and treatment / editor, Werner Zimmerli. Other titles: Bone and joint infections (Zimmerli) Description: Second edition. | Hoboken, NJ, USA : Wiley-Blackwell, 2021. | Includes bibliographical references and index. | Description based on print version record and CIP data provided by publisher; resource not viewed. Identifiers: LCCN 2020047588 (print) | LCCN 2020047589 (ebook) | ISBN 9781119720683 (epub) | ISBN 9781119720669 (adobe pdf) | ISBN 9781119720652 (hardback) | ISBN 9781119720652q(hardback) | ISBN 9781119720669q(adobe pdf) | ISBN 9781119720683q(epub) Subjects: | MESH: Osteomyelitis | Arthritis, Infectious Classification: LCC RC931.O7 (ebook) | LCC RC931.O7 (print) | NLM WE 251 | DDC 616.7/15–dc23 LC record available at https://lccn.loc.gov/2020047588 LC record available at https://lccn.loc.gov/2020047589 Cover Design: Wiley Cover Image: Courtesy of Werner Zimmerli Set in 10/11.5pts Times New Roman MT Std by SPi Global, Pondicherry, India 10 9 8 7 6 5 4 3 2 1

Contents

List of Contributors Preface to the Second Edition Foreword to the First Edition Acknowledgments Chapter 1 Introduction Werner Zimmerli

xii xvi xvii xix 1

Chapter 2  Diagnostic Approach in Bone and Joint Infections 5 Nora Renz, Donara Margaryan, and Andrej Trampuz Introduction 5 Common Microorganisms Causing Bone and Joint Infection 7 Diagnostic Approach in Spinal Infection 8 Diagnostic Approach in Bone Fixation Device‐Associated Infection 10 Diagnostic Approach in Native Arthritis and Infections after Anterior Cruciate Ligament Repair (ACL‐R) 12 Diagnostic Approach in Periprosthetic Joint Infections (PJI) 15 Key Points 17 References 17 Chapter 3  Unusual Microorganisms in Periprosthetic Joint Infection 21 Camelia Marculescu and Werner Zimmerli Introduction 21 Gram‐Positive Microorganisms 21 Gram‐Negative Microorganisms 26 Zoonotic Microorganisms 29 Anaerobic Microorganisms 34 Mycobacteria 35 Other Microorganisms 38 Key Points 40 References 40 v

vi

Contents

Chapter 4  Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques 51 Maria Eugenia Portillo and Stéphane Corvec Introduction 51 Broad‐Range PCR 51 Targeted PCR 53 Multiplex PCR 54 Next‐Generation Sequencing Approach 55 Mass Spectrometry‐Based Methods 58 Key Points 61 References 61 Chapter 5  Bacteriophages for Treatment of Biofilm Infections 65 Mercedes Gonzalez‐Moreno, Paula Morovic, Tamta Tkhilaishvili, and Andrej Trampuz History of Bacteriophage Use in Human Infections 65 Principles of Bacteriophage Therapy 66 Activity of Bacteriophages against Bacterial Biofilms and Persisters 69 Bacteriophage Susceptibility Testing: the Phagogram 71 Experimental and Clinical Evidence with Bacteriophage Treatment 72 Local Delivery and Systemic Bacteriophage Application 74 Outlook and Future Perspectives 76 Key Points 77 References 78 Chapter 6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 81 Cornelia B. Landersdorfer, Jürgen B. Bulitta, Roger L. Nation, and Fritz Sörgel Pharmacokinetics 81 Bone Sample Preparation and Analysis 82 Pharmacokinetic Sampling and Data Analysis 83 Penetration of Antibiotics into Bone 84 Pharmacodynamics and Monte Carlo Simulations 93 Conclusions 95 Key Points 96 References 96 Chapter 7  Preclinical Models of Infection in Bone and Joint Surgery 99 Caroline Constant, Lorenzo Calabro, Willem‐Jan Metsemakers, R. Geoff Richards, and T. Fintan Moriarty Introduction 99 Influence of Species in Preclinical Models of Bone and Joint Infection 100 Overview of Animal Models 103 Direct Inoculation with Minimal Trauma 103 Animal Models of Bone and Joint Infection Incorporating Trauma 108 Hematogenous Models 108 Future Directions 109 Conclusions 110 Key Points 110 References 110

Contents

vii

Chapter 8  Native Joint Arthritis in Children 117 Pablo Yagupsky Introduction 117 Epidemiology 117 Microbiology 118 Pathogenesis 122 Clinical Presentation 123 Laboratory Investigation 125 Imaging Studies 127 Differential Diagnosis 128 Treatment 128 Prognosis 133 Key Points 134 References 134 Chapter 9  Native Joint Arthritis in Adults 139 Florian B. Imhoff, David E. Bauer, and Ilker Uçkay Introduction 139 Risk Factors 139 Pathogenesis, Epidemiology, and Microbiology 140 Diagnosis 141 Treatment 144 Outcome 147 Key Points 148 Acknowledgments 148 References 148 Chapter 10  Septic Arthritis of Axial Joints 151 Werner Zimmerli Septic Arthritis of the Sternoclavicular Joint 151 Key Points 157 Septic Arthritis of the Symphysis Pubis 157 Key Points 161 Septic Arthritis of the Sacroiliac Joint 161 Key Points 165 References 165 Chapter 11  Periprosthetic Joint Infection: General Aspects 171 Werner Zimmerli Introduction 171 Definition 172 Classification 172 Pathogenesis 172 Laboratory Investigation 176 Therapeutic Management 178 Prophylaxis 179 Errors in the Management of PJI 180 Key Points 180 References 181

viii

Contents

Chapter 12  Periprosthetic Joint Infection after Total Hip and  Knee Arthroplasty 187 Werner Zimmerli, Rihard Trebse, and Martin Clauss Introduction 187 Risk Factors 188 Microbiology 188 Clinical Features 189 Laboratory Investigation 190 Imaging Procedures 191 Management 192 Instructive Cases 202 References 206 Chapter 13  Periprosthetic Joint Infection after Shoulder Arthroplasty 213 Parham Sendi, Andreas Marc Müller, Beat K. Moor, and Matthias A. Zumstein Introduction 213 Risk Factors 213 Microbiology 215 Pathogenesis 216 Clinical Features 217 Laboratory Investigation 218 Imaging Procedures 220 Management 220 Instructive Cases 224 References 226 Chapter 14  Periprosthetic Joint Infection after Elbow Arthroplasty 231 Yvonne Achermann, Michael C. Glanzmann, and Christoph Spormann Introduction 231 Microbiology 232 Clinical Features 232 Diagnostic Procedures 234 Management 237 Instructive Cases 242 References 245 Chapter 15  Periprosthetic Joint Infection after Ankle Arthroplasty 251 Parham Sendi, Bernhard Kessler, and Markus Knupp Introduction 251 Risk Factors 252 Microbiology 254 Clinical Features 254 Laboratory Investigation 255 Imaging Procedures 256 Management 257 Instructive Cases 260 References 262

Contents

ix

Chapter 16  Osteomyelitis: Classification 265 Werner Zimmerli Classification According to Pathogenesis 265 Classification According to the Duration of Infection 267 Classification According to the Localization 268 Classification According to the Presence of an Implant 269 Classification According to Anatomy and Comorbidity 269 References 270 Chapter 17  Osteomyelitis in Children 273 Alexander Aarvold, Priya Sukhtankar, and Saul N. Faust Introduction 273 Epidemiology 273 Pathophysiology 274 Clinical Presentation, Diagnosis, and Microbiology 275 Treatment 282 Complications 284 Key Points 285 References 286 Chapter 18  Acute Osteomyelitis in Adults 289 Werner Zimmerli Introduction 289 Pathogenesis 289 Epidemiology 290 Microbiology 291 Risk Factors 293 Clinical Features 295 Laboratory Investigation 296 Imaging Procedures 298 Clinical and Imaging Differential Diagnosis 300 Treatment 301 Key Points 304 References 304 Chapter 19  Subacute Osteomyelitis: Tuberculous and Brucellar Vertebral Osteomyelitis 309 Juan D. Colmenero and Pilar Morata Introduction 309 Epidemiology 309 Clinical Features 310 Laboratory Investigation 312 Imaging Procedures 315 Antimicrobial and Surgical Therapy 318 Key Points 321 References 321

x

Contents

Chapter 20  Chronic Osteomyelitis in Adults 325 Felix W.A. Waibel, Benedikt Jochum, and Ilker Uçkay Introduction 325 Pathogenesis 326 Diagnosis 326 Treatment 328 Key Points 333 Acknowledgments 333 References 333 Chapter 21  Diabetic Foot Osteomyelitis 337 Eric Senneville and Olivier Robineau Introduction 337 Classification 337 Microbiology 339 Risk Factors 341 Clinical Features 342 Inflammatory Parameters 342 Imaging Procedures 343 Treatment 344 Prevention 347 References 348 Chapter 22  Osteomyelitis of the Jaws 353 Werner Zimmerli Introduction 353 Classification 353 Microbiology 355 Risk Factors 356 Clinical Features 356 Laboratory Investigation 359 Imaging Procedures 360 Management 361 Key Points 363 References 364 Chapter 23  Fracture‐Related Infection of the Long Bones 367 Parham Sendi, Mario Morgenstern, Willem‐Jan Metsemakers, and Martin McNally Introduction 367 Classification and Risk Factors 368 Microbiology 370 Clinical Features 371 Laboratory Investigation 372 Imaging Procedures 373 Management 373 Instructive Cases 380 References 383

Contents

xi

Chapter 24  Implant‐Associated Vertebral Osteomyelitis 387 Todd J. Kowalski and Arick P. Sabin Introduction 387 Classification and Risk Factors 388 Microbiology 389 Clinical Features 393 Diagnostic Procedures 393 Treatment 395 Instructive Cases 402 References 404 Chapter 25  Postoperative Sternal Osteomyelitis 409 Parham Sendi, Mihai Constantinescu, and Lars Englberger Introduction 409 Risk Factors 410 Microbiology 414 Clinical Features 414 Laboratory Investigation 415 Imaging Procedures 417 Management 417 Instructive Cases 422 References 425 Index

429

List of Contributors

Alexander Aarvold, MD Paediatric Orthopaedics University Hospital Southampton NHS Foundation Trust Southampton, UK Yvonne Achermann, MD Division of Infectious Diseases and Hospital Epidemiology University Hospital Zürich University of Zürich Zürich, Switzerland David E. Bauer, MD Department of Orthopedic Surgery Balgrist University Hospital University of Zürich Zürich, Switzerland Jürgen B. Bulitta, PhD Department of Pharmacotherapy and Translational Research College of Pharmacy University of Florida Orlando, FL, USA Lorenzo Calabro, MD QEII Jubilee Hospital Brisbane Brisbane, Queensland, Australia xii

Martin Clauss, MD Centre for Musculoskeletal Infections University Hospital Basel Department of Orthopaedic and Trauma Surgery University of Basel Basel, Switzerland Juan D. Colmenero, MD Infectious Diseases Service University Regional Hospital Malaga, Spain Caroline Constant, DMV, MSc, MENG, DACV‐LA AO Research Institute Davos AO Foundation Davos, Switzerland Mihai Constantinescu, MD Department of Plastic Reconstructive, and Hand Surgery University Hospital of Bern University of Bern Bern, Switzerland Stéphane Corvec, PharmD, PhD Clinical Microbiology Service de Bactériologie‐Hygiène hospitalière Institut de Biologie ‐ CHU de Nantes Nantes, France

List of Contributors

Lars Englberger, MD Department of Cardiac Surgery Hirslanden Clinic Aarau and Bern Bern, Switzerland Saul N. Faust, MD NIHR Southampton Clinical Research Facility University Hospital Southampton NHS Foundation Trust Faculty of Medicine and Institute for Life Sciences University of Southampton Southampton, UK Mercedes Gonzalez‐Moreno, MSc Charité‐Universitätsmedizin Corporate member of Freie Universität Berlin Humboldt‐ Universität zu Berlin, and Berlin Institute of Health Center for Musculoskeletal Surgery Berlin, Germany Florian B. Imhoff, MD Department of Orthopedic Surgery Balgrist University Hospital University of Zurich Zürich, Switzerland Benedikt Jochum, MD Department of Orthopedic Surgery Balgrist University Hospital University of Zurich Zürich, Switzerland Bernhard Kessler, MD Internal Medicine and Infectious Diseases Hospital Emmental, Burgdorf Burgdorf, Switzerland Marcus Knupp, MD Mein Fusszentrum University of Basel Basel, Switzerland

xiii

Todd J. Kowalski, MD Department of Internal Medicine Gundersen Health System University of Wisconsin School of Medicine and Public Health La Crosse, WI, USA Cornelia B. Landersdorfer, PhD Centre for Medicine Use and Safety Monash Institute of Pharmaceutical Sciences Monash University (Parkville Campus) Melbourne, Australia Camelia Marculescu, MD Division of Infectious Diseases Department of Medicine Medical University of South Carolina Charleston, SC, USA Donara Margaryan, MD Charité‐Universitätsmedizin Corporate Member of Freie Universität Berlin Humboldt‐ Universität zu Berlin, and Berlin Institute of Health Center for Musculoskeletal Surgery Berlin, Germany Martin McNally, MD The Bone Infection Unit Nuffield Orthopaedic Centre Oxford University Hospitals Oxford, UK Willem‐Jan Metsemakers, MD Department of Trauma Surgery University Hospitals Leuven Department of Development and Regeneration, KU Leuven Leuven, Belgium Beat K. Moor, MD Service of Traumatology and Orthopedic Surgery Wallis Hospital Center Martigny, Switzerland

xiv

List of Contributors

Pilar Morata, PhD Biochemistry and Molecular Biology Faculty of Medicine University of Malaga Malaga, Spain T. Fintan Moriarty, PhD AO Research Institute Davos AO Foundation Davos, Switzerland Mario Morgenstern, MD Centre for Musculoskeletal Infections University Hospital Basel Department of Orthopaedic and Trauma Surgery University of Basel Basel, Switzerland Paula Morovic Charité‐Universitätsmedizin Corporate Member of Freie Universität Berlin Humboldt‐ Universität zu Berlin, and Berlin Institute of Health Center for Musculoskeletal Surgery Berlin, Germany Andreas Marc Müller, MD Centre for Musculoskeletal Infections University Hospital Basel Department of Orthopaedic and Trauma Surgery University of Basel Basel, Switzerland Roger L. Nation, PhD Drug Delivery Disposition and Dynamics Monash Institute of Pharmaceutical Sciences Monash University (Parkville Campus) Melbourne, Australia Maria Eugenia Portillo, MD, PhD Clinical Microbiology Complejo Hospitalario de Navarra Irunlarrea Pamplona, Navarra, Spain

Nora Renz, MD Charité‐Universitätsmedizin Corporate member of Freie Universität Berlin Humboldt‐ Universität zu Berlin, and Berlin Institute of Health Center for Musculoskeletal Surgery Berlin, Germany R. Geoff Richards, PhD AO Research Institute Davos AO Foundation Davos, Switzerland Olivier Robineau, MD Infectious Diseases Department University Hospitals of Tourcoing Lille University Tourcoing, France Arick P. Sabin, MD Department of Internal Medicine Gundersen Health System La Crosse, WI, USA Parham Sendi, MD Centre for Musculoskeletal Infections University Hospital Basel Institute for Infectious Diseases University of Bern Bern, Switzerland Eric Senneville, MD, PhD Infectious Diseases Department University Hospitals of Tourcoing Lille University Tourcoing, France Fritz Sörgel, PhD IBMP‐Institute for Biomedical and Pharmaceutical Research Nürnberg‐Heroldsberg, and Institute of Pharmacology Faculty of Medicine University of Duisburg‐Essen Essen, Germany

List of Contributors

Christoph Spormann, MD Upper Extremities Hirslanden Clinic and Endoclinic Zürich Zürich, Switzerland Priya Sukhtankar NIHR Southampton Clinical Research Facility University Hospital Southampton NHS Foundation Trust Southampton, UK Tamta Tkhilaishvili, MD Charité‐Universitätsmedizin Corporate member of Freie Universität Berlin Humboldt‐ Universität zu Berlin, and Berlin Institute of Health Center for Musculoskeletal Surgery Berlin, Germany Andrej Trampuz, MD Charité‐Universitätsmedizin Corporate member of Freie Universität Berlin Humboldt‐ Universität zu Berlin, and Berlin Institute of Health Center for Musculoskeletal Surgery Berlin, Germany Rihard Trebse, MD Valdoltra Orthopaedic Hospital Ankaran University of Ljubljana Medical Faculty Ljubljana, Slovenija

xv

Ilker Uçkay, MD Department of Orthopedic Surgery Infectiology Balgrist University Hospital University of Zürich Zürich, Switzerland Felix W.A. Waibel, MD Department of Orthopedic Surgery Balgrist University Hospital University of Zürich Zürich, Switzerland Pablo Yagupsky, MD Clinical Microbiology Laboratory Soroka University Medical Center Ben Gurion University of the Negev Beer Sheva, Israel Werner Zimmerli, MD Infectious Diseases Interdisciplinary Unit for Orthopedic Infections Kantonsspital Baselland University of Basel Liestal, Switzerland Matthias A. Zumstein, MD Shoulder, Elbow, and Orthopedic Sports Medicine Orthopaedics Sonnenhof, Bern Bern, Switzerland

Preface to the Second Edition

The knowledge in the field of bone and joint infections is steadily increasing. Looking at PubMed® shows that the number of publications in the field of septic arthritis, periprosthetic joint infection, and osteomyelitis has been strongly rising for 8–15 years. Thus, it seems natural that a new edition of this textbook dealing with bone and joint infections would facilitate the daily clinical decisions of specialized and non‐specialized physicians. Due to an unstoppable specialization of infectious diseases, a textbook with information weighted by dedicated specialists in the field is still useful, despite the unlimited availability of information in the internet. The main drawback of unfiltered publications in medical databases is the necessity of the appropriate selection by the clinician. For this purpose, specialized clinical experience is needed. Therefore, a textbook written by recognized experts in their field allows to make the correct clinical decisions. In the 2nd edition, all chapters have been updated to the current scientific knowledge. In addition, four new chapters have been added. Since the correct diagnosis is the base of the rational treatment, a chapter with diagnostic algorithms for the most important topics in the field of bone and joint infections is added. Implants increase the susceptibility to infection by compromising the local host defense. Therefore, literally every microorganism may cause periprosthetic joint infection. A new chapter presents clinical information about such rare infections caused by unusual bacteria. In routine diagnostic labs, many microbiologists may lack specialized knowledge in the field of non‐culture techniques. Therefore, a new chapter presents these methods for the application in patients with bone and joint infection. The correct use of these methods is discussed. This allows the rational use of non‐culture techniques without wasting laboratory capacity and money. Finally, the fourth new chapter is dedicated to bacteriophages, which have been recently used in the treatment of different chronic infections also in western countries. This therapeutic technique is of increasing interest, due to the rising problem of multiresistant bacteria. Up to now, controlled clinical trials have been missing. The presentation of the current knowledge may motivate clinicians to plan such studies in patients with biofilm infections. I hope that the new edition, written by a dedicated multidisciplinary team of specialists, is a useful guide to the rational management of patients with bone and joint infections. Werner Zimmerli, MD xvi

Foreword to the First Edition

Correct and rapid diagnosis and treatment of bone and joint infections require the ­collaboration of many different specialists. In the field of bone and joint infections, there is almost no evidence based on controlled trials. Therefore, learning from the clinical experience of experts is particularly important for the management of such infections. Up to now, a comprehensive, internationally available textbook dealing with infections of the locomotor system has been missing. The present book fills this gap. An editor of a book on bone and joint infections must have a broad knowledge of the pertinent literature. In addition, he or she should know the different competent specialists in the field. When John Wiley & Sons asked Werner Zimmerli to edit this book, they could not have come up with a better choice. He is a competent specialist in the field of infections of the locomotor system and also has contacts with a network of international specialists. Werner Zimmerli started his professional career in infectious diseases in Geneva, in the group of Francis A. Waldvogel, a well‐known specialist of osteomyelitis, whose pioneer publications are still widely cited. Under his leadership, Werner Zimmerli started his ­studies on implant‐related infections. In the early 1980s, it was rather an exception for an infectious disease specialist to invade a field for which orthopedic surgeons were responsible. Together with Daniel Lew and Pierre Vaudaux, he was able to define host factors that are responsible for the high susceptibility of implants to pyogenic infections. In subsequent studies in Basel, he presented experimental evidence that implants are susceptible not only for exogenous but also for hematogenous infections. Together with Andreas F. Widmer, he could show the special role of rifampin for the treatment of implant‐associated infections in vitro and in vivo. In conjunction with my group in Liestal, we started to treat patients with orthopedic implant‐associated staphylococcal infections with rifampin combinations. The promising treatment results in observational studies led to the planning of a randomized controlled trial on the role of rifampin in patients with implant retention. In this trial, rifampin showed its superiority in patients suffering from acute orthopedic implant‐associated infections treated with debridement and implant retention. After the planned interim analysis, the trial was stopped early, because all treatment failures were observed in the arm without rifampin. This study offers one of the few evidence‐based treatment standards in orthopedic infections. xvii

xviii

Foreword to the First Edition

As an orthopedic surgeon specializing in the field of infections of the locomotor system, I met Werner Zimmerli, a partner who is a dedicated clinician. Together with my team in Liestal, I was in permanent contact with him to discuss our cases with bone and joint infections. This contact was even closer when he accepted the position of Head of the Basel University Medical Clinic in Liestal. This allowed us to create an “Interdisciplinary Unit for Orthopedic Infections,” the first in Switzerland. Out of this close collaboration resulted a now‐internationally respected algorithm for the treatment of periprosthetic joint infections. Over the last two decades, a large number of Infectious Disease specialists and orthopedic surgeons trained in the field of bone and joint infections in our group. The interdisciplinary core team later included plastic surgeons, microbiologists, and pathologists, a concept of collaboration that is now widely accepted. This book reflects the concepts of interdisciplinarity. The introductory chapters deal with general fields, which are important for managing bone and joint infections, namely “Microbiology,” “Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone,” and “Experimental Preclinical Models.” In addition, the book contains general chapters on “Periprosthetic Joint Infections” and “Classification of Osteomyelitis.” These chapters offer to the reader detailed knowledge as a basis for competent clinical management of bone and joint infections. The main part of the book deals with the typical clinical entities of infections of the locomotor system. They are written following a common concept and contain all the knowledge needed to better understand the subject treated. Each chapter can be studied independently. Most of the chapters end with an enumeration of key points and some instructive cases illustrating typical situations. This allows the reader to assess whether he or she understood the essentials of management. Several case examples also illustrate common errors that should be avoided. Extensive and updated lists of references help the reader continue his or her studies. The chapters are written by specialists with extensive clinical experience in the field of the infections that have been described. If the treatment is mainly conservative, an infectious disease specialist is the author. In the chapters on infections associated with prosthetic joints and ­internal fixation devices, an orthopedic surgeon joined the infectious disease specialist in the writing team. This book offers clear information on most problems in the management of infections of the locomotor system. Experienced clinicians in the fields of infectious diseases, orthopedic surgery, trauma surgery, rheumatology, and internal medicine can use this book as a comprehensive textbook or on a chapter‐by‐chapter basis. Specialists in the field can benefit from the detailed updated reviews and may find specific help for their own challenging cases. Peter E. Ochsner Professor Emeritus in Orthopedic Surgery of the University of Basel

Acknowledgments

I am grateful to Dr. Julia Squarr (Senior Commissioning Editor) for her support in the planning of the new edition, as well as to Rosie Hayden, Managing Editor from John Wiley & Sons, for her continuous help during the realization of this project. My thanks go also to Emma Cole for her careful and efficient copyediting. This textbook would not have been possible without the enthusiastic and competent work of the authors, who are all specialists in different fields of bone and joint infection. My special thanks go to Ruth Mester, who was indispensable during the whole editorial process. I would also like to thank all the anonymous patients whose bone and joint infections were the basis of new diagnostic and therapeutic concepts.

xix

Chapter 1

Introduction Werner Zimmerli

The prevalence of most bone and joint infections is steadily increasing, mainly due to the rising life expectancy of the population, and the increased use of bone fixation devices and prosthetic joints. For frequent infectious diseases, such as respiratory tract, urinary tract, and bloodstream infections, many diagnostic and therapeutic aspects have been studied in a controlled fashion [e.g. 1–3]. In contrast, in the field of bone and joint infections, randomized controlled trials are rare. Exceptions are a randomized controlled study on the role of rifampin in patients with orthopedic implant‐associated infections, and a controlled trial comparing two different durations of antibiotic treatment in patients with vertebral osteomyelitis [4,5]. Therefore, diagnostic and therapeutic advice must be based mainly on individual clinical expert knowledge and observational studies [6–10]. The optimal diagnostic and therapeutic management of bone and joint infections needs special know‐how in different fields of medicine. Many physicians have only limited clinical experience, since arthritis and osteomyelitis are rare infectious diseases. Therefore, a multidisciplinary approach to these infections is desirable. Only for a few topics are internationally accepted guidelines for the management of bone and joint infections available [11–13]. In addition, publications on the clinical practice comprising different aspects of these infections are scarce. The aim of this book is to close this gap with texts from a multidisciplinary team of experts in the field. Indeed, specialists in microbiology, clinical pharmacology, preclinical research, pediatrics, pediatric and adult orthopedic surgery, infectious diseases, and cardiovascular surgery contributed to this book. This broad spectrum of expertise made it possible to cover a wide range of pathophysiological, epidemiological, diagnostic, and therapeutic aspects of bone and joint infection. The principal focus of the book is on clinical practice. It should enable clinicians to manage patients according to the best available evidence. Besides the routine microbiological tests, novel non‐culture techniques are increasingly used for the diagnosis of infectious diseases, including bone and joint infection. However, the clinical role of molecular diagnostic procedures and mass spectrometry is ill defined. Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

1

2

Bone and Joint Infections

The potential advantages of these techniques are a more rapid identification and a higher sensitivity, especially in patients with antibiotic pretreatment or with difficult‐to‐detect microorganisms [14,15]. With a worldwide increase in multidrug resistance, alternative antimicrobial therapies are looked for. The use of bacteriophages is a promising option in patients with bone and joint infection caused by multiresistant bacteria. Bacteriophages have a long history, however it is only recently that experimental and clinical data appeared in the literature [16]. In the near future, controlled clinical trials will show their role in biofilm infections. The role of the bone/serum ratio in the antimicrobial treatment of bone and joint infections is still a matter of debate. Important methodological differences have to be considered to adequately judge data on bone penetration. These data are often controversially discussed in the literature, mainly due to the use of varying experimental techniques in different studies [17,18]. Distinct differences in the extent of bone penetration by various classes of antimicrobial agents have been observed. However, the proof for the clinical relevance of these differences is still missing. Thus, knowledge about pharmacokinetics and pharmacodynamics of antibiotics in bone should stimulate planning of clinical studies to fill this missing gap. Many current treatment concepts are based on preclinical studies in vitro and in animals [19]. Such data are especially important for the management of implant‐associated infections, a field in which controlled clinical trials are lacking. Septic arthritis encompasses a non‐homogenous group of joint infections. In this book, eight different clinical situations are covered. In arthritis of children, many aspects differ from arthritis in adults. In children, Kingella kingae plays a prominent role, a microorganism which in adults almost exclusively causes endocarditis [20]. In addition, Streptococcus agalactiae is still common in neonates. In contrast, Haemophilus influenzae type b almost disappeared in young children due to the effective conjugate vaccine. Arthritis of axial joints is rare and difficult to diagnose. IV‐drug use is the most frequent risk factor of all types of axial arthritis, namely of the sternoclavicular joint, the symphysis pubis, and the sacroiliac joint. Surgery is rarely needed if the diagnosis is rapidly made and the patient has no pyogenic complications. Prosthetic joints are increasingly used not only in hip and knee, but also in other joints, mainly shoulder, ankle, and elbow. The perioperative infection rate ranges from about 0.5–1.5% after hip or knee arthroplasty up to 10% after elbow or ankle joint replacement. Since many aspects vary between the ­different joint prostheses, separate chapters deal with periprosthetic joint infection in this book. Osteomyelitis encompasses a large spectrum of different diseases. Many different ­classifications are used, depending on different aspects of disease (e.g. pathogenesis, duration, presence of implant) and according to the specialist who is managing the case (e.g. orthopedic surgeon, infectious disease specialist, pediatrician, angiologist). In this book, aspects of age (children, adults), duration of disease (acute, subacute, chronic), presence of implant, anatomic location (long bones, vertebrae, jaws), and presence of diabetes are presented in separate chapters. Together with all authors, I trust that this multidisciplinary book will allow the gathering of rapid and exhaustive information regarding all types of bone and joint infection. If this book allows you to improve patient management, we have reached our goal.

1 Introduction

3

References 1. Torres A, Zhong N, Pachl J, et al. Ceftazidime‐avibactam versus meropenem in nosocomial pneumonia, including ventilator‐associated pneumonia (REPROVE): a ­randomised, double‐blind, phase 3 non‐inferiority trial. Lancet Infect Dis. 2018;18(3): 285–295. 2. Wagenlehner FME, Cloutier DJ, Komirenko AS, et al. Once‐daily plazomicin for complicated urinary tract infections. N Engl J Med. 2019;380(8):729–740. 3. Wirz Y, Meier MA, Bouadma L, et al. Effect of procalcitonin‐guided antibiotic treatment on clinical outcomes in intensive care unit patients with infection and sepsis patients: a patient‐ level meta‐analysis of randomized trials. Crit Care. 2018;22(1):191. 4. Zimmerli W, Widmer AF, Blatter M, et  al. Role of rifampin for treatment of orthopedic implant‐related staphylococcal infections: a randomized controlled trial. Foreign‐Body Infection (FBI) Study Group. JAMA. 1998;279(19):1537–1541. 5. Bernard L, Dinh A, Ghout I, et al. Antibiotic treatment for 6 weeks versus 12 weeks in patients with pyogenic vertebral osteomyelitis: an open‐label, non‐inferiority, randomised, controlled trial. Lancet. 2015;385(9971):875–882. 6. Gellert M, Hardt S, Koder K, et al. Biofilm‐active antibiotic treatment improved the outcome of knee periprosthetic joint infection: Results from a 6‐year prospective cohort. Int J Antimicrob Agents. 2020:105904. 7. Roux S, Valour F, Karsenty J, et al. Daptomycin > 6 mg/kg/day as salvage therapy in patients with complex bone and joint infection: cohort study in a regional reference center. BMC Infect Dis. 2016;16:83. 8. Lowik CAM, Parvizi J, Jutte PC, et  al. Debridement, antibiotics and implant retention is a viable treatment option for early periprosthetic joint infection presenting more than four weeks after index arthroplasty. Clin Infect Dis. 2019. 9. Depypere M, Morgenstern M, Kuehl R, et al. Pathogenesis and management of f­ racture‐ related infection. Clin Microbiol Infect 2019; https://doi.org/10.1016/j.cmi. 2019.08.006. 10. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic‐joint infections. N Engl J Med. 2004;351(16):1645–1654. 11. Osmon DR, Berbari EF, Berendt AR, et  al. Diagnosis and management of prosthetic joint infection: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis. 2013;56(1):e1–e25. 12. Berbari EF, Kanj SS, Kowalski TJ, et al. Infectious Diseases Society of America (IDSA) clinical practice guidelines for the diagnosis and treatment of native vertebral osteomyelitis in adults. Clin Infect Dis. 2015;61(6):e26–46. 13. Lipsky BA, Berendt AR, Cornia PB, et  al. Infectious Diseases Society of America clinical practice guideline for the diagnosis and treatment of diabetic foot infections. Clin Infect Dis. 2012;54(12):e132–173. 14. Street TL, Sanderson ND, Atkins BL, et al. Molecular diagnosis of orthopedic‐device‐related infection directly from sonication fluid by metagenomic sequencing. J Clin Microbiol. 2017;55(8):2334–2347. 15. Thoendel MJ, Jeraldo PR, Greenwood‐Quaintance KE, et al. Identification of prosthetic joint infection pathogens using a shotgun metagenomics approach. Clin Infect Dis. 2018;67(9):1333–1338. 16. Tkhilaishvili T, Winkler T, Muller M, et al. Bacteriophages as adjuvant to antibiotics for the treatment of periprosthetic joint infection caused by multidrug‐resistant pseudomonas aeruginosa. Antimicrob Agents Chemother. 2019;64(1). 17. Mouton JW, Theuretzbacher U, Craig WA, et  al. Tissue concentrations: do we ever learn? J Antimicrob Chemother. 2008;61(2):235–237.

4

Bone and Joint Infections

18. Landersdorfer CB, Bulitta JB, Kinzig M, et al. Penetration of antibacterials into bone: pharmacokinetic, pharmacodynamic and bioanalytical considerations. Clin Pharmacokinet. 2009;48(2):89–124. 19. Vanvelk N, Morgenstern M, Moriarty TF, et al. Preclinical in vivo models of fracture‐related infection: a systematic review and critical appraisal. Eur Cell Mater. 2018;36:184–199. 20. Yagupsky P. Kingella kingae: from medical rarity to an emerging paediatric pathogen. Lancet Infect Dis. 2004;4(6):358–367.

Chapter 2

Diagnostic Approach in Bone and Joint Infections Nora Renz, Donara Margaryan, and Andrej Trampuz

Introduction A meticulous diagnostic workup is paramount for a successful management of bone and joint infections, as the surgical and medical treatment differ considerably between aseptic and septic entities. Since low‐virulent pathogens present with only subtle clinical signs and symptoms of infection, the diagnosis represents a challenge. It requires a comprehensive approach involving several diagnostic tests, especially in the presence of an implant. Once the infection is diagnosed, additional infection features determine the treatment strategy, including duration (acuity) of infection, source of infection (pathogenesis), and expected or isolated pathogen(s). Precise history taking, clinical examination, and imaging help to determine the duration of infection. The acuity of infection is particularly relevant in implant‐associated infections, as the biofilm age (“maturity”) guides the surgical management. In acute infections, the implant can generally be successfully retained, whereas in chronic infections, it should be removed or exchanged. Similarly, the duration of infection directs the surgical strategy in osteomyelitis. While in acute osteomyelitis without implant, surgery is generally dispensable, in chronic osteomyelitis surgical debridement with removal of all dead material (sequestrectomy) is required. Clues for the pathogen are the acuity of symptoms, pathogenesis, previous episodes of bone or joint infection, and the local epidemiology [1]. Because source control is required to improve the treatment outcome, identification of the hematogenous route of infection is of the utmost importance and allows to extend diagnostic measures in order to search and identify the primary focus of infection (Table 2.1). Depending on the primary focus of infection, the treatment strategy needs to be adapted, i.e. the intravenous treatment prolonged to four to six weeks, or the primary infection source controlled by surgical intervention. If the primary infection focus is not

Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

5

6

Bone and Joint Infections

Table 2.1.  Investigation of cause in case of hematogenous bone and joint infections. Pathogen

Focus

Investigation

Staphylococcus aureus

• Skin • Infective endocarditis • Intravascular devices or catheters • Primary bacteremia

• Skin examination • Check for intravascular implants • Blood cultures (BC) • Transesophageal echocardiography (TEE) in case of positive BC

Coagulase‐ negative staphylococci

• Intravascular devices or catheters • Infective endocarditis

• Check for intravascular implants • Blood cultures • TEE in case of positive BC

Viridans group (Streptococcus mitis/oralis)

• Oral cavity • Infective endocarditis

• Check for recent dental procedure • Orthopantomogram • Blood cultures • TEE in case of positive BC

S. agalactiae, S. dysgalactiae

• Abdomen • Urogenital tract • Skin

• Skin examination • Urinalysis and culture • Imaging of abdomen/ pelvis in case urinalysis is normal

S. gallolyticus

• Colorectal cancer or adenoma • Infective endocarditis

• Blood cultures • TEE in case of positive BC • Colonoscopy (if not done recently)

Gram‐ negative rods

Escherichia coli, Klebsiella spp., Enterobacter spp., Pseudomonas spp.

• Abdomen • Urogenital tract

• Urinalysis and culture • Imaging of abdomen/ pelvis (colonoscopy) if not performed recently and urinalysis is normal

Enterococcus spp.

E. faecalis, E. faecium, other spp.

• Abdomen • Urogenital tract • Infective endocarditis

• Urinalysis and culture • Blood cultures • TEE in case of positive blood cultures • Colonoscopy if not performed recently and urinalysis is normal

Staphylo‐ coccus spp.

Streptococcus spp.

identified and treated accordingly, a secondary infection may reoccur due to relapsing spread from the primary focus, despite correct treatment of the bone or joint infection [2]. The preoperative diagnostic methods are generally less sensitive and specific than intraoperative methods. Therefore, preoperative diagnostic tests, such as imaging

2  Diagnostic Approach in Bone and Joint Infections

7

and synovial fluid analysis, need to be re‐evaluated and interpreted together with intraoperative findings. Microbiological and histopathological analysis of tissue samples, as well as sonication fluid cultures of removed foreign bodies, may allow improvement of the etiologic diagnosis. In unclear cases in the preoperative setting, one must plan the treatment presuming infection and avoid insufficient surgical debridement or partial retention of implant material.

Common Microorganisms Causing Bone and Joint Infection Most pathogens causing bone and joint infections originate from the skin or from mucosal surfaces of the oral cavity, urogenital, and intestinal tract, i.e. from the patient´s microbiome. The majority of bone and joint infections is caused by Gram‐ positive cocci, mostly by Staphylococcus spp., which account for 50–60% of all infections [3]. The remaining percentages are comprised of Streptococcus spp., Enterococcus spp., Gram‐positive and Gram‐negative anaerobes, and other rare pathogens such as fungi, mycobacteria, or intracellular bacteria  [4]. After open fractures with dirty wounds, environmental microorganisms are involved, whereas in unintentional use of contaminated medical devices (breach of sterility), typically nosocomial pathogens are observed. The proportion of individual pathogens depends on the pathogenesis of infection, the anatomic location, and the time of occurrence in case of postoperative infections  [3]. Whereas native septic arthritis, vertebral osteomyelitis without previous instrumentation, and acute osteomyelitis in children are predominantly caused by hematogenous seeding into the affected bone or joint, only approximately 30% of periprosthetic joint infections (PJI) are caused hematogenously. In infections associated with implants, such as fracture‐fixation devices or spine instrumentation hardware, hematogenous infection is extremely rare (1 cm)?

yes

no

Repeat biopsy

Open or CT-guided biopsy of the spine/disc1 Histopathology Microbiology

no

BC negative

yes

Surgery with intraoperative diagnostics3 Histopathology Microbiology +/– sonication

Hematogenous infection: search for distant focus: TEE (vegetation?) OPG Intravascular device (vascular catheter, CIED, prosthetic valve) Urinalysis X-ray of lung

Contiguous infection: search for adjacent focus: Examination of skin (cellulitis/ ulcer?) Imaging of abdomen/pelvis (abscess?)

Histopathology and/or microbiology2 consistent with infection? yes

Conservative treatment with antimicrobial therapy

No clinical improvement/ progressive mechanical instability?

yes

1

If blood cultures grow the causing pathogen in clinically and radiologically proven vertebral osteomyelitis (without implant), biopsy of spine or disc may be omitted. For highly virulent organisms (e.g. Staphylococcus aureus, Escherichia coli, Streptococcus spp.) or for patients on antibiotic therapy one positive sample confirms infection, for low-virulent organisms (e.g. Staphylococcus epidermidis, Cutibacterium acnes) ≥2 positive samples are required to confrim infection. 3 Additional microbiological investigations (Mycobacterium spp., Brucella spp.), if exposure/risk factors present. BC: blood cultures, TEE: transesophageal echocardiography, OPG: orthopantomogram, CIED: cardiac implantable electronic device 2

Figure 2.1.  Diagnostic algorithm for spinal infections (with and without implant).

New onset back pain or neurologic deficits, accompanied by systemic inflammatory signs and/or wound healing disturbances, should prompt a diagnostic assessment. Before spine surgery, history of fever/rigors, clinical examination, determination of systemic inflammatory markers, and imaging of the spine should be performed (conventional

10

Bone and Joint Infections

X‐ray, CT, or MRI, as appropriate [see Chapter 24]). In case of new onset of back pain without prior intervention (i.e. surgery or infiltration/injections) or in febrile patients, blood cultures should be collected before antimicrobial therapy. Depending on the ­isolated pathogen, further investigation of the primary source of infection should be initiated in parallel to the investigation of the spine. Native spine infections are mostly of hematogenous origin and require collection of blood cultures. Transcutaneous tissue sampling (in general CT‐guided biopsy) is needed, if blood cultures remain negative and conditions requiring surgery are absent. Most cases with hematogenous infection can be treated with antibiotics alone [14] (see Chapter 18). Therefore, comprehensive microbiological sampling before initiation of antimicrobial treatment is crucial. In contrast, implant‐associated spinal infections predominantly occur after perioperative contamination. Surgical debridement is always needed, and the diagnosis is confirmed with intraoperative diagnostic tests (see Chapter 24). During surgery, multiple tissue samples (three to five samples for microbiology and histopathology) are collected and removed hardware (either entire internal fixation material or only loose implants) should be sent for sonication fluid culture. Carlson et al. [15] showed sonication of removed spinal implants to be more sensitive than peri‐implant tissue ­culture, reaching statistical significance when using a threshold of ≥20 CFU/10mL for sonication culture positivity. The removed implant should be sent to sonication to exclude low‐grade infection in all spinal surgeries performed for presumed mechanical (aseptic) complications such as implant loosening, which may manifest several years after instrumentation. This observation is supported by Prinz et al. [16] who demonstrated microbial growth in 22 of 82 patients (41%) with screw loosening, as compared to none of 28 patients with firm screws (p 10,000 cells/μL is 90% and a CRP >100 mg/L is 77%, but both parameters are nonspecific at this threshold [24]. If the patient is septic or suffers from fever or rigors, immediate collection of blood cultures and arthrocentesis followed by initiation of empiric treatment is paramount. Septic arthritis is a medical emergency due to the rapid cartilage damage by ­leukocyte proteases. Therefore, prompt arthrocentesis  –  preferably before antibiotic ­treatment  –  is the first step in the assessment of a painful inflamed joint, even in patients without systemic signs of infection. Synovial fluid examination consists of analysis of leukocyte count and percentage of granulocytes (use EDTA tubes to prevent clumping), Gram stain, conventional culture (inoculate preferentially pediatric blood culture bottles to increase culture sensitivity), and crystals by polarized light microscopy (for details see Chapter 9). Gram stain has a low sensitivity. However, due to the high specificity, it can be used as a rule‐in test in case of a positive result. The presence of crystals in synovial fluid does not rule out infection, as the combination of crystal-induced and septic arthritis may occur [25]. If any of the analyses in synovial fluid is consistent with infection, immediate surgical intervention (arthroscopy or arthrotomy) is indicated. Intraoperatively, the diagnosis is completed with histopathological and microbiological analysis of synovial membrane. Furthermore, the patient should be examined for primary infection foci, including intravascular, skin and soft‐tissue, urogenital, gastrointestinal, or pulmonary infections, if no intervention or adjacent infectious focus at the site of the joint preceded the acute onset of symptoms. Novel biomarkers in synovial fluid such as (D‐lactate, interleukin‐6 (IL‐6), total lactate D‐ and L‐isomers), and calprotectin have been investigated for the diagnosis of septic arthritis in recent years. Among them, mainly D‐lactate and calprotectin showed a promising performance (sensitivity of 85% and 76%, respectively, and specificity of 96 and 94%, respectively) [26,27]. Interleukin‐6, however, did not allow for a reliable differentiation between septic and aseptic arthritis [28]. Infections after ACL‐R represent a combination of a native joint arthritis with transplant or implants in situ (avascular graft and fixation devices). In most cases, concomitant osteomyelitis of the adjacent bone, where the new graft is attached with fixation devices, is present. The most common pathogenesis is perioperative contamination of the graft or joint. Most infections occur in the first few weeks after reconstruction and present with acute onset of signs and symptoms of infection [29]. Clinical distinction between physiological postoperative process and signs of infection is difficult. Similarly, the interpretation of synovial fluid leukocyte count is challenging and cut‐off values for diagnosing infection in this setting have not been determined yet. This is especially true in chronic low‐grade infections caused by low‐virulent pathogens, evoking little local or systemic inflammation. Nevertheless, synovial leukocyte count of >10,000/μL may ­indicate septic arthritis, whereas normal counts exclude it. In case of instable transplant, exchange of graft/fixation device is required, and the removed parts should be used for microbiological analysis, i.e. conventional culture of the graft and sonication of the fixation devices [29].

2  Diagnostic Approach in Bone and Joint Infections

15

Diagnostic Approach in Periprosthetic Joint Infections (PJI) The diagnosis of acute PJI is straightforward, because the clinical context is often obvious, and most diagnostic tests have a high sensitivity. In contrast, chronic low‐grade infections are difficult to differentiate from aseptic prosthetic failures. Therefore, a comprehensive algorithmic approach combining preoperative and intraoperative results is required to reliably confirm or exclude infection (Figure 2.4). Every painful prosthetic joint should be assessed for infection. Initial examinations include clinical examination of the patient, assessment of systemic inflammatory parameters, and imaging. Systemic inflammatory parameters are neither sensitive nor specific for the diagnosis of PJI (see Chapter 11). Nevertheless, they represent a relevant puzzle piece in the diagnostic workup. As in native joint infection, patients presenting with sepsis or with fever/rigors should be promptly assessed for hematogenous infection, and blood and synovial fluid should be collected, followed by immediate empiric antimicrobial treatment. Revision surgery should take place within a few hours after admission in order to realize source control. However, septic revision of a prosthetic joint is a challenging procedure requiring special expertise. Therefore, it should be performed by an experienced orthopedic surgeon. The focus lies on intraoperative diagnostics such as histopathological and microbiological analysis including synovial fluid analysis (culture and leukocyte count), periprosthetic tissue (at least three to five samples) and sonication of the retrieved implant parts [30] (see Chapter 11). Depending on the pathogen and the clinical presentation, investigation of the primary focus with additional diagnostic tests should be performed. In late acute onset infections with negative blood cultures, the possibility of contiguous infection by direct expansion of a nearby focus onto the joint should be considered. In chronic infection, the initial focus lies on the preoperative assessment. The infection should ideally be diagnosed or excluded before revision, which allows the planning of the most appropriate treatment strategy. In the preoperative setting, purulent wound secretion and/or sinus tract communicating with the prosthetic joint are confirmative signs of PJI. In this case no further diagnostic measures are needed, and revision surgery with intraoperative diagnostics should be scheduled. In the absence of confirmatory clinical signs, the most important and sensitive diagnostic measure in the preoperative setting is arthrocentesis. It should be performed according to standard aseptic technique, preferably in the operating room. Culture of synovial fluid has a low sensitivity, as only planktonic bacteria are detected, and bacteria embedded in the biofilm on the implant surface remain unrecognized. Synovial fluid leukocyte count analysis is more sensitive, as it reflects the host reaction against the microorganisms. However, in situations associated with aseptic inflammatory changes of the joint, the specificity is compromised, including the healing process in the first four to six weeks after surgery, inflammatory changes after trauma, recurrent dislocations, and in underlying inflammatory arthropathies. Furthermore, the optimal cutoff for the diagnosis of infection is subject to debate. Generally, at this stage of the diagnostic algorithm, a test with a high sensitivity is preferred, therefore lower leukocyte count (e.g. 2000/μL) is preferred, risking overdiagnosis rather than missing an infection. The performance of novel biomarkers such as synovial fluid alpha defensin, D‐lactate, calprotectin, and leukocyte esterase have been assessed applying different definition criteria, however leukocyte count showed equal or better performance [31]. In case of dry joint tap, instillation of saline is not recommended, as the leukocyte count analysis is no longer representative due to dilution of synovial fluid. If the synovial fluid

16

Bone and Joint Infections

(vascular catheter, CIED, prosthetic valve)

implantable electronic device

Figure 2.4.  Diagnostic algorithm for periprosthetic joint infection.

, CIED: cardiac

2  Diagnostic Approach in Bone and Joint Infections

17

analysis is inconclusive or inconsistent with infection, differential diagnoses should be considered and further investigated. If the pain remains unexplained, further diagnostic procedures depend on the level of suffering of the patient and the stability of the prosthesis. In case of severe pain with significant compromise of life quality or in case of a loosened prosthesis, revision surgery (preferably one‐stage revision) with meticulous intraoperative diagnostics is advocated. Biopsy through arthroscopy or mini‐open arthrotomy, depending on the affected joint and the expertise of the treating orthopedic surgeon, is an alternative in case of a stable prosthesis. However, the diagnostic yield is rarely significant, as the representative area (i.e. interface between implant and bone) is not reached. In selected cases, diagnostic exchange of mobile parts with consecutive sonication fluid culture may be indicated. Nevertheless, every intervention represents a risk for superinfection, and should therefore be avoided. If the pain is under adequate control with analgesics, and the patient does not urge for immediate intervention, a repeated arthrocentesis after one to three months should be considered. Intraoperative diagnostics including periprosthetic tissue histopathology and microbiology as well as sonication fluid culture of the removed implants are more sensitive than preoperative arthrocentesis. Therefore, these procedures are nowadays considered the gold standard in the diagnosis of PJI [19,30,32]. Details regarding these diagnostic methods are discussed in Chapter 11. In delayed and late PJI and in polymicrobial infections, as well as in PJI caused by low‐virulent pathogens, the detection rate is significantly higher in intraoperatively collected specimens compared to synovial fluid culture [33].

Key Points ●●

●●

●●

●●

The diagnostic workup of bone and joint infections aims at identifying the pathogenesis and acuity of the infection, as well as the responsible pathogen, in order to plan an adequate treatment strategy. Most common pathogens causing bone and joint infections originate from the ­patient’s skin and mucosal microbiome; their proportion vary depending on the route of infection. Accurate intraoperative tests are histopathological and microbiological examination including synovial fluid, periprosthetic tissue, and sonication fluid culture. In case of septic patients with bone and joint infections, blood cultures should be collected before initiation of antimicrobial treatment in order to diagnose primary or, less frequently, secondary bacteremia. In case of suspected hematogenous periprosthetic joint infection, the isolated pathogen guides the rational further diagnostic workup to detect the primary focus.

References 1. Zeller V, Kerroumi Y, Meyssonnier V, et al. Analysis of postoperative and hematogenous prosthetic joint‐infection microbiological patterns in a large cohort. J Infect. 2018;76(4): 328–334. 2. Rakow A, Perka C, Trampuz A, et al. Origin and characteristics of haematogenous periprosthetic joint infection. Clin Microbiol Infect. 2019;25(7):845–850. 3. Tande AJ, Patel R. Prosthetic joint infection. Clin Microbiol Rev. 2014;27(2):302–345.

18

Bone and Joint Infections

4. Corvec S, Portillo ME, Pasticci BM, et al. Epidemiology and new developments in the diagnosis of prosthetic joint infection. Int J Artif Organs. 2012;35(10):923–934. 5. Koder K, Hardt S, Gellert MS, et al. Outcome of spinal implant‐associated infections treated with or without biofilm‐active antibiotics: results from a 10‐year cohort study. Infection. 2020. (Epub ahead of print) 6. Margaryan D, Renz N, Bervar M, et al. Spinal implant‐associated infections: A prospective multicenter cohort study. Int J Antimicrob Agents. 2020:106116. 7. Widmer AF, Frei R, Rajacic Z, et al. Correlation between in vivo and in vitro efficacy of antimicrobial agents against foreign body infections. J Infect Dis. 1990;162(1):96–102. 8. Zimmerli W. Experimental models in the investigation of device‐related infections. J Antimicrob Chemother. 1993;31 Suppl D:97–102. 9. Lowik CAM, Zijlstra WP, Knobben BAS, et al. Obese patients have higher rates of polymicrobial and Gram‐negative early periprosthetic joint infections of the hip than non‐obese patients. PLoS One. 2019;14(4):e0215035. 10. Ohl CA, Forster D. Infectious arthritis of native joints. In: Bennett JE, Dolin R, Blaser MJ, et al. Mandell, Douglas, and Bennett’s Principles and Practice of Infectious Diseases. 8th edition, pp 1302–1317, ed. Philadelphia, PA: Elsevier/Saunders; 2015. 11. Mylona E, Samarkos M, Kakalou E, et  al. Pyogenic vertebral osteomyelitis: a systematic review of clinical characteristics. Semin Arthritis Rheum. 2009;39(1):10–17. 12. Yusuf E, Steinrucken J, Buchegger T, et al. A descriptive study on the surgery and the microbiology of Gustilo type III fractures in an university hospital in Switzerland. Acta Orthop Belg. 2015;81(2):327–332. 13. Giesecke MT, Schwabe P, Wichlas F, et al. Impact of high prevalence of pseudomonas and polymicrobial Gram‐negative infections in major sub‐/total traumatic amputations on empiric antimicrobial therapy: a retrospective study. World J Emerg Surg. 2014;9(1):55. 14. Zimmerli W. Clinical practice. Vertebral osteomyelitis. N Engl J Med. 2010;362(11): 1022–1029. 15. Sampedro MF, Huddleston PM, Piper KE, et  al. A biofilm approach to detect bacteria on removed spinal implants. Spine. 2010;35(12):1218–1224. 16. Prinz V, Bayerl S, Renz N, et al. High frequency of low‐virulent microorganisms detected by sonication of pedicle screws: a potential cause for implant failure. J Neurosurg Spine. 2019;31(3):424–429. 17. Rutges JPHJ, Kempen DH, van Dijk M, et al. Outcome of conservative and surgical treatment of pyogenic spondylodiscitis: a systematic literature review. Eur Spine J. 2015; 25(4):983–999. 18. Berbari EF, Kanj SS, Kowalski TJ, et al. Executive Summary: 2015 Infectious Diseases Society of America (IDSA) Clinical Practice Guidelines for the Diagnosis and Treatment of Native Vertebral Osteomyelitis in Adults. Clin Infect Dis. 2015;61(6):859–863. 19. Onsea J, Depypere M, Govaert G, et al. Accuracy of tissue and sonication fluid sampling for the diagnosis of fracture‐related infection: A Systematic Review and Critical Appraisal. J Bone Jt Infect. 2018;3(4):173–181. 20. Metsemakers WJ, Morgenstern M, McNally MA, et al. Fracture‐related infection: A consensus on definition from an international expert group. Injury. 2018;49(3):505–510. 21. Govaert GAM, Kuehl R, Atkins BL, et al. Diagnosing fracture‐related infection: Current concepts and recommendations. J Orthop Trauma. 2020;34(1):8–17. 22. Depypere M, Morgenstern M, Kuehl R, et  al. Pathogenesis and management of fracture‐ related infection. Clin Microbiol Infect. 2020;26(5):572–578. 23. Morgenstern M, Athanasou NA, Ferguson JY, et al. The value of quantitative histology in the diagnosis of fracture‐related infection. Bone Joint J. 2018;100‐b(7):966–972. 24. Conen A, Borens O. Septic Arthritis. In: Kates SL, Borens O, editors. Principles of Orthopedic Infection Management. Davos, Switzerland: AO Publishing; 2016. p. 213–226. 25. Shah K, Spear J, Nathanson LA, et al. Does the presence of crystal arthritis rule out septic arthritis? J Emerg Med. 2007;32(1):23–26.

2  Diagnostic Approach in Bone and Joint Infections

19

26. Baillet A, Trocme C, Romand X, et al. Calprotectin discriminates septic arthritis from pseudogout and rheumatoid arthritis. Rheumatology (Oxford, England). 2019;58(9):1644–1648. 27. Gratacos J, Vila J, Moya F, et al. D‐lactic acid in synovial fluid. A rapid diagnostic test for bacterial synovitis. J Rheumatol. 1995;22(8):1504–1508. 28. Lenski M, Scherer MA. Analysis of synovial inflammatory markers to differ infectious from gouty arthritis. Clin Biochem. 2014;47(1–2):49–55. 29. Mouzopoulos G, Fotopoulos VC, Tzurbakis M. Septic knee arthritis following ACL reconstruction: a systematic review. Knee Surg Sports Traumatol Arthrosc. 2009;17(9):1033–1042. 30. Dudareva M, Barrett L, Figtree M, et al. Sonication versus tissue sampling for diagnosis of prosthetic joint and other orthopedic device‐related infections. J Clin Microbiol. 2018;56(12). 31. Renz N, Yermak K., Perka C., Trampuz A. Alpha defensin lateral flow test for diagnosis of periprosthetic joint infection. Not a screening but a confirmatory test. J Bone Joint Surg Am. 2018(100(9)):742–750. 32. Krenn V, Morawietz L, Perino G, et al. Revised histopathological consensus classification of joint implant related pathology. Pathol Res Pract. 2014;210(12):779–786. 33. Schulz P, Dlaska CE, Perka C, et al. Preoperative synovial fluid culture poorly predicts the pathogen causing periprosthetic joint infection. Infection. 2020 Nov 3. doi: 10.1007/s15010020-01540-2. Epub ahead of print.

Chapter 3

Unusual Microorganisms in Periprosthetic Joint Infection Camelia Marculescu and Werner Zimmerli

Introduction The association of certain microorganisms such as Staphylococcus epidermidis, Staphylococcus aureus, and ß‐hemolytic streptococci with periprosthetic joint infection (PJI) has been recognized for many years. These microorganisms are usually easily ­isolated on culture using conventional media. Other microorganisms that can cause PJI may be more difficult to culture or are not typically associated with PJI. Recent advances in molecular microbiology and culture procedures have led to the discovery of microorganisms that are less commonly associated with PJI. The recognition of PJI due to these microorganisms is important because the course of untreated PJI is consistently unfavorable.

Gram‐Positive Microorganisms Staphylococcus lugdunensis S. lugdunensis is a coagulase‐negative staphylococcus (CNS) whose pathogenic potential is similar to that of S. aureus. It also develops a biofilm phenotype which may appear as small‐ colony variant [1]. Although genetically indistinguishable, small‐colony variants differ in size and antibiotic susceptibility from the parent strain, and are responsible for chronic persistent infections and failure of antibiotic treatment. Several cases of PJIs after total knee (TKA) and total hip arthroplasty (THA) due to S. lugdunensis have been reported, in both immunocompromised (rheumatoid arthritis, multiple myeloma, p ­ ancreatic cancer, diabetes mellitus) and immunocompetent patients. The microorganism more often causes PJI after TKA than THA. Presentation of PJIs due to Staphylococcus lugdunensis can be acute, with fever and local signs of inflammation or persistent pain at the surgical site. Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

21

22

Bone and Joint Infections

Microbiologic Diagnosis S. lugdunensis frequently produces a clumping factor that may result in positive slide (short) coagulase test. Therefore, identification of S. lugdunensis should include tube (long) coagulase test. Additional biochemical tests (positive results on pyrrolidonyl‐ arylamidase and ornithine‐decarboxylase tests) are helpful for final identification. Therapy Therapy for S. lugdunensis PJI included two‐stage exchange (TSE), debridement and retention of implant (DAIR), and one‐stage exchange (OSE). There are insufficient data to determine whether the outcome of S. lugdunensis PJI treated with DAIR will be ­similar to the outcome of S. aureus PJI [1–4].

Streptococcus bovis The association between S. bovis bacteremia and the presence of an underlying colonic pathology (colorectal carcinoma, polyps, colonic ulcers) has been extensively reported [5,6]. Some rare cases of S. bovis PJIs were reported in the literature [7,8]. Emerton et  al. [8] reported one case of S. bovis THA PJI occurring one month after a nonspecific flu‐like ­illness, in the absence of endocarditis. Microbiologic Diagnosis Colonies of S. bovis are typically non‐hemolytic on sheep blood agar. S. bovis grows in 40% bile and hydrolyzes esculin. It can be distinguished from enterococci because the former is PYR negative, fails to hydrolyze arginine, and is unable to grow in 6.5% salt broth. S. bovis can be further divided into biotype I and II, based on additional biochemical reactions. 16S rRNA sequencing allows further differentiation of biotype II strains into groups 1 and 2 [9]. The microorganism is usually susceptible to penicillin. Therapy Antimicrobial therapy without revision surgery was unsuccessful in two cases with S.  bovis PJI, as could be expected. Subsequently, both patients underwent TSE with favorable outcome after 74 and 6 months of follow‐up, respectively [7,8]. In the series of Thompson et  al. [6], patients with acute PJI were treated with DAIR, and those with chronic PJI with TSE. Both patients treated with DAIR and two patients with TSE got suppressive therapy for unknown reasons. In the five episodes managed with TSE, treatment was successful. A six‐week course of ceftriaxone was found to be effective [6]. A  majority of the reported cases had colonic pathology. Thus, colonic investigation should be pursued for patients diagnosed with S. bovis PJI.

Gemella G. morbillorum (formerly Streptococcus morbillorum), G. haemolysans, and G. sanguinis PJI have been reported [10–12]. G. morbillorum is rarely involved in human infections such as endocarditis. G. sanguinis was reported in one case as a late acute infection [12].

3  Unusual Microorganisms in Periprosthetic Joint Infection

23

Microbiologic Diagnosis G. morbillorum is an anaerobic to aerotolerant Gram‐positive coccus. Specimens ­collected at the time of surgery should be inoculated on aerobe and anaerobe plates. Improving the rate of isolation of such fastidious microorganisms from the joint fluid may require direct inoculation of joint fluid in blood culture bottles [12,13]. G. morbillorum is identified as a Gram‐positive, pleomorphic, slow‐growing, and initially anaerobic bacteria. Identification of G. morbillorum may be impeded by its slow growth rate, as well as variable morphology and staining properties. Identification of G. haemolysans is usually established by a series of biochemical reactions (API 20 Strep identification system) together with Gram stain morphology. Partial 16s rRNA sequencing identified G. sanguinis [12]. Better aerobic growth allows differentiation of G. haemolysans from viridans group streptococci and G. morbillorum. Gemellae are typically susceptible to penicillin, erythromycin, tetracycline, and vancomycin. Therapy Reported cases were treated with TSE or DAIR. One patient with G. haemolysans TKA PJI was treated with resection arthroplasty, six weeks of parenteral penicillin G, followed by reimplantation with favorable outcome after two months of follow‐up [10]. Chronic oral suppression with doxycycline was used in the case treated with DAIR [12].

Listeria monocytogenes L. monocytogenes is a Gram‐positive, nonspore‐forming aerobic rod that causes a variety of infections in humans. PJI due to Listeria is very rare. The source of listeriosis is usually unknown, but a number of reports have implicated the consumption of unpasteurized milk/cheese, vegetables, or processed meat. L. monocytogenes PJI tend to occur in elderly patients or immunocompromised patients with malignancy, transplantation, diabetes, cirrhosis or rheumatoid arthritis. PJI caused by Listeria spp. typically presents as late hematogenous infection. Microbiologic Diagnosis Diagnosis is based on Gram stain and culture of the microorganism. Listeria produces a characteristic appearance on sheep blood agar with small zones of clear beta hemolysis around each colony. In addition, saline suspensions of Listeria grown in vitro demonstrate characteristic tumbling motility, whereas Corynebacterium spp. (ie. diphtheroids) do not exhibit motility. Listeria grows well at refrigeration temperatures (4° to 10°C). Blood cultures can be negative in cases of PJI [14, 15]. In vitro, Listeria is susceptible to ampicillin, trimethoprim‐sulfamethoxazole, aminoglycosides, and vancomycin. Cephalosporins are not effective against Listeria and false positive in vitro sensitivity reports can result from the use of disk diffusion tests [15,16]. Therapy Prosthesis removal was the surgical modality that eventually led to cure of the infection in five cases. DAIR and OSE followed by antimicrobial suppression were also reported

24

Bone and Joint Infections

with good short‐term outcomes [15,17,18]. Failure of conservative management ­requiring OSE was recently reported after eight months of follow‐up. This patient had no signs of infection after one year of follow‐up [19]. The recommended medical therapy is ­parenteral ampicillin with or without gentamicin. Gentamicin should not be used as monotherapy [16,17]. Little information about the optimal duration of treatment is available. Six weeks of intravenous antimicrobials appear to be adequate in reported cases. Relapses can occur because of the involvement of foreign material and the intracellular growth of the microorganism. The duration of chronic oral antimicrobial suppression in cases treated with prosthesis retention needs to be assessed carefully, especially in immunocompromised patients. Profoundly immunocompromised patients may require life‐long oral ­antimicrobial suppression.

Nocardia Nocardia ssp. are opportunistic microorganisms, causing infection mainly in immunocompromised patients (cancer, hematological malignancies, particularly lymphoid, chronic respiratory diseases, corticosteroid treatment, transplant recipients, sarcoidosis). Several cases of Nocardia (nova, asteroids, and cyriacigeorgica) THA and TKA PJIs have been reported, two in immunocompetent patients, and the remainder in immunocompromised patients [20–23]. Nocardia nova presented as a late, indolent infection in a patient with systemic lupus erythematosus, two years after being diagnosed with pulmonary nocardiosis. In an immunocompetent patient, Nocardia PJI was presumably caused perioperative contamination [21]. Microbiologic Diagnosis Diagnosis is based on the typical morphology on the modified acid‐fast stain. It appears as a thin, Gram‐positive microorganism with branching filaments with a beaded appearance. The acid‐fast reaction and the production of aerial branching allow differentiation of Nocardia from other aerobic and anaerobic actinomycetes. When deep surgical cultures are obtained, they are usually positive 85–90% of the time [21,24]. However, even when adequate specimens are obtained, recovery of Nocardia in the laboratory can be difficult. Most routine bacterial, fungal, or mycobacterial culture media can support the growth of Nocardia. However, Nocardia requires prolonged incubation of up to two to three weeks. Speciation of the different taxons of Nocardia is difficult. The optimal method for antimicrobial susceptibility testing has not been established, and there are no specific NCCLS MIC breakpoints for Nocardia [21,25]. It is therefore recommended to send the Nocardia isolates to a reference laboratory. In the reported cases of PJI, N. nova was susceptible to imipenem, erythromycin, and amikacin. In contrast to N. asteroides, N.  nova is resistant to trimethoprim‐sulfamethoxazole. Although N. nova may demonstrate susceptibility to ampicillin or amoxicillin in vitro, it consistently carries a beta‐­ lactamase, which may hydrolyze these antibiotics [21,25]. Therapy The drugs of choice for the treatment of N. asteroides infections are either sulfonamides or tetracyclines. The suggested duration of therapy to prevent relapses for cutaneous and pulmonary disease is three months, particularly with trimethoprim‐sulfamethoxazole ­

3  Unusual Microorganisms in Periprosthetic Joint Infection

25

c­ombinations [21,26]. However, the efficacy of combinations remains controversial, ­particularly against N. nova. Imipenem‐amikacin combination appears to be more effective than ­ trimethoprim‐sulfamethoxazole in nocardiosis, in particular N. nova infection. Linezolid is active in vitro against all Nocardia spp. TSE and even OSE followed by chronic oral suppression have yielded favorable outcomes in the reported cases [20–23].

Dietzia maris D. maris is an environmental actinomycete that is rarely involved in human disease. To date, only one case of THA infection has been reported. D. maris was presumably acquired as superinfection during a prolonged free interval with a spacer [27]. Microbiologic Diagnosis Diagnosis is based on the Gram‐stain from intraoperative specimens that disclosed many Gram‐positive cocci germinating into short rods. The isolate was identified as D. maris using the API Coryne strip. Rapid molecular methods and cellular fatty acid analysis confirmed the identification of D. maris. The microorganism was susceptible, by disk‐­ diffusion method, to amoxicillin, imipenem, gentamicin, trimethoprim‐sulfamethoxazole, rifampin, clindamycin, vancomycin, and pristinamycin. Therapy This patient was treated with teicoplanin for four months without further surgery. No follow‐up data were reported.

Tsukamurella T. paurometabolum is a Gram‐positive, weakly or variably acid‐fast, nonmotile, obligate aerobic bacillus that exists primarily as a saprophyte in the soil. Tsukamurella should be recognized as a potential pathogen in patients with immunosuppression, indwelling foreign bodies, and postoperative wounds. In one case, Tsukamurella was found after repeated debridements and removal of a TKA for a mixed (Peptostreptococcus and CNS) infection [28]. Microbiologic Diagnosis Diagnosis was based on bone culture results that yielded Gram‐positive rods after 13 days of incubation. Final identification was done five weeks later. The microorganism is susceptible to sulfamethoxazole, clarithromycin, imipenem, amikacin, ciprofloxacin, rifampin, vancomycin, and third‐generation cephalosporins. Therapy The reported case was treated with TSE and a two‐month course of clarithromycin plus ciprofloxacin. Reimplantation was performed four months after explantation. The outcome was successful, however, without mentioning the duration of follow‐up.

26

Bone and Joint Infections

Oerskovia Oerskovia are Gram‐positive, Nocardia‐like bacilli which inhabit the soil and rarely cause human infections. One case of late, presumably exogenous TKA‐PJI with O. xanthineolytica, has been reported [29]. The patient suffered from intermittent pain and swelling after a knee trauma. Revision arthroplasty was performed for presumed aseptic loosening, but O. xanthineolytica was diagnosed at removal of the implant. Microbiologic Diagnosis Diagnosis was based on the Gram stain and culture of the surgical specimens, as well as biochemical tests using API Coryne strip. Gram stain revealed Gram‐positive branching diphtheroid rods. Susceptibility testing for many of these fastidious, slow growing microorganisms is not yet standardized and differences in the MIC of the microorganism with different method used might be encountered [29]. The most active antimicrobial in vitro is vancomycin [29,30]. Penicillin or ampicillin, rifampin, and vancomycin are drugs of choice in infections caused by this microorganism. Therapy The patient was treated with TSE and a five‐week course of trimethoprim‐sulfamethoxazole. Reimplantation was performed three months after explantation. The outcome was ­excellent after six months of follow‐up.

Bacillus alvei Infections with non‐anthrax‐Bacillus spp. generally occur in immunocompromised patients or intravenous drug users. A 26‐year‐old woman with sickle‐cell disease suffered from B. alvei sepsis seeding on a THA 12 years after implantation. The patient was treated with removal of the prosthesis and a six‐week course of vancomycin. The patient remained well after 18 months of follow‐up [31].

Bacillus cereus B. cereus is a catalase‐positive, aerobic, spore‐forming bacillus. Soft tissue and bone infections due to this pathogen have been associated with trauma, intravenous drug use, and immunodeficiency. B. cereus isolated from blood has been easily dismissed as contaminant without repeated isolation from multiple blood cultures. A THA PJI due to Bacillus cereus and bacteremia occurring 13 years after THA has been reported in a diabetic female [32]. The patient was treated with implant removal followed by intravenous vancomycin. The outcome cannot be judged due to the very short follow‐up.

Gram‐Negative Microorganisms Salmonella Nontyphoidal Salmonella infection is mainly seen in young or debilitated patients, in patients with sickle cell disease, collagen vascular disease, immunosuppressive medications, or HIV. Among reported cases with PJI, mainly after THA, one patient had sickle

3  Unusual Microorganisms in Periprosthetic Joint Infection

27

cell trait and others had different immunocompromising conditions such as rheumatoid arthritis, renal transplantation, familial Mediterranean fever, ankylosing spondylitis, immunosuppressive meds, chronic lymphatic leukemia, ulcerative colitis, lung adenocarcinoma, and polymyositis [33–40]. S. typhimurium was the most frequently isolate seen. Other reported Salmonella species to cause PJI were S. dublin, newport, muenchen, hirschfeldii, enteritidis, and choleraesuis. Salmonella PJI is typically hematogenously seeded from the gastrointestinal tract. The presentation is acute, usually secondary to bacteremia and/or gastroenteritis. Infections may either occur in the early or late postoperative period. Microbiologic Diagnosis Salmonella are relatively easy to identify in the clinical microbiology laboratory. They grow under both aerobic and anaerobic conditions. Salmonella are oxidase and lactose‐ negative (clear to semitranslucent colonies on MacConkey agar plates). S. typhi, S. paratyphi C, and S. dublin strains have the Vi antigen. A DNA sequence encoding the Vi antigen was used in developing a nested PCR for S. typhi [41]. In one study it was noted that the prevalence of S. typhimurium isolates with resistance to ampicillin, chloramphenicol, streptomycin, sulfonamides, and tetracycline was 34% in 1996 [42]. In 2001, 60% of S. cholearaesuis isolates were resistant to ciprofloxacin [43]. Given the increasing rate of resistance among Salmonella strains, therapy should be guided by the in vitro susceptibility studies. For susceptible isolates, fluoroquinolones may be the preferred agent, since they have activity against biofilm bacteria [46]. Therapy The preferred surgical therapy is removal of prosthesis followed by antimicrobial therapy. Day et al. [44] reported a case of a TKA infection due to Salmonella enteritidis that was treated with DAIR and exchange of mobile parts, followed by six weeks of ceftriaxone. Chronic oral antimicrobial suppression was not used. In this case the outcome was excellent after six years of follow‐up. In another case, a patient with S. dublin THA infection was also treated with DAIR. Trimethoprim‐sulfamethoxazole was administered for two years. The patient did not show any clinical signs of recurrence [45]. One case of S. dublin THA PJI relapsed after OSE. Ciprofloxacin was administered for one year, and the infection was cured after three and a half years of follow‐up [46]. TSE for Salmonella THA infection was associated with an excellent outcome in two patients [47,48]. In three cases no surgical intervention was done, and patients were maintained on chronic oral suppression [49–51]. The infection relapsed in one patient. As could be expected, antimicrobial therapy alone is insufficient to cure PJI.

Neisseria meningitidis Primary meningococcal arthritis is an unusual form of disseminated meningococcal disease in which the features of acute pyogenic arthritis develop without meningitis or meningococcemia. Three cases of PJI due to Neisseria meningitides were reported in the literature. All three involved a TKA and were treated with DAIR [52,53,54]. Vikram et al. [52] described an 80‐year‐old woman without obvious risk factors for PJI or meningococcal disease. She suffered from primary meningococcal TKA PJI. The onset of symptoms was acute, and the patient had associated meningococcal bacteremia, without evidence of meningitis.

28

Bone and Joint Infections

Microbiologic Diagnosis Neisseria meningitidis will grow both on sheep blood agar and chocolate agar or chocolate agar containing antibiotics (e.g. Thayer‐Martin agar). N. gonorrhoeae does not grow on sheep blood agar. Both N meningitidis and N. gonorrhoeae utilize glucose. N. meningitidis is differentiated from N. gonorrhoeae by utilization of maltose. Therapy All patients were treated with DAIR, followed by variable courses of parenteral antimicrobial therapy (ranging 3–6 weeks), followed by a short course of oral amoxicillin or ciprofloxacin for six to nine weeks [53,54], or indefinite chronic oral suppression with penicillin [52].

Haemophilus spp. There are sporadic reported cases of Haemophilus influenzae THA and TKA PJIs [55,56]. Risk factors for Haemophilus influenzae arthritis include multiple myeloma, systemic lupus erythematous, rheumatoid arthritis, chronic lymphatic leukemia, common variable hypogammaglobulinemia, diabetes mellitus, and alcohol abuse. A depressed immune ­status may play a role in the development of such infections. Blood cultures are usually positive. All reported cases were acute hematogenous infections. Microbiologic Diagnosis Diagnosis is based on the Gram stain from a joint aspirate showing Gram‐negative coccobacilli. H. influenzae grows aerobically as pinpoint colonies on chocolate blood agar. Speciation is done by the differential growth requirements for hematin (X factor) and nicotinamide factor (V factor). Therapy Patients were treated with DAIR, since they had no evidence of prosthesis loosening and the infections were acute. Patients received intravenous antimicrobial therapy. An oral fluoroquinolone was used for suppression during one year in one case [57]. One patient required resection arthroplasty, followed by six weeks of IV‐ampicillin and long‐term oral amoxicillin [58]. The outcome of H. influenzae PJI was favorable, after a follow‐up between six months and five years. Routine vaccination against Haemophilus influenza serotype B may be helpful in patients with prosthetic joints and hematological malignancies predisposing to bacterial infections [56]. A patient with two prosthetic hip joints suffered from bilateral PJI caused by Haemophilus parainfluenza after extraction of several loosening teeth without prophylaxis [59]. Microbiologic diagnosis relies on Gram‐stain, aerobic culture of the microorganism on chocolate agar and determination of X and V factor requirements for growth. H. parainfluenzae is usually susceptible to ampicillin, cefuroxime, ceftriaxone, quinolones and trimethoprim‐sulfamethoxazole. Septic loosening occurred after three months in one case, and required TSE for cure [59]. Another patient with TKA PJI due to H. parainfluenzae after dental s­ urgery

3  Unusual Microorganisms in Periprosthetic Joint Infection

29

improved after an eight‐week course of intravenous ceftriaxone, followed by oral antimicrobials (trimethoprim‐sulfamethoxazole and subsequently oral ciprofloxacin) for two years [60].

Zoonotic Microorganisms Brucella spp. PJI due to Brucella is infrequent. Brucella melitensis infection in humans is transmitted from animals through contaminated dairy products, nonpasteurized milk, by contact through a skin laceration, or by inhalation. To date, 35 cases of Brucella PJI have been reported [61–66]. Almost all patients consumed unpasteurized dairy products or had occupational exposure (farmers, contact with cattle or goats, wild meat butchers). Brucella PJI presents as an indolent infection, mainly with local symptoms. Systemic symptoms were present in only 13/34 (38%). One patient had a sinus tract [64]. The median time from prosthesis implantation to diagnosis of PJI was 48 months (range between 0 and 168 months). More than half of the patients had documented loosening on radiographs. Brucella melitensis was isolated in the vast majority of patients (25/35), followed by Brucella abortus (4/35), suis (1 case), and Brucella spp. (5/35). Microbiologic Diagnosis Diagnosis was mostly made by positive cultures in joint fluid, rarely by blood cultures, or intraoperative tissue biopsies. Coinfection with other pathogens is reported. A negative joint culture result cannot rule out osteoarticular brucellosis. Using the BACTEC blood culture system (Becton Dickinson), detection of brucellae can be accomplished with a sensitivity of 95% within 6 days [67]. Blood PCR‐ELISA for Brucella is more sensitive (94.9%) and specific (96.5%) than conventional manual blood culture systems. Diagnosis of Brucella spp. infection is based on identification of Gram‐negative, oxidase‐positive cocci and rods. Once suspicious colonies are isolated, differentiation of Brucella species is usually accomplished by agglutination with specific antiserum. An agglutination test of the serum and/or the ­synovial fluid to determine Brucella antibodies with titer greater than 1:160 is considered evidence for the presence of Brucella infection [65]. Synovial α‐defensin test was reported negative in a case of documented bacteremic Brucella melitensis bilateral TKA PJI [68]. Therapy Management of Brucella osteomyelitis and PJI is controversial with regard to antimicrobial selection, duration of therapy, and the role of surgery. A recent meta‐analysis showed that combination therapy with doxycycline/aminoglycoside (streptomycin or gentamicin) is preferred over doxycycline/rifampin or other combinations [69]. The optimal treatment duration is unknown, ranging from 6 weeks to 19 months in the current series. Malizos and al. [65] suggest longer duration of therapy of patients with joint implants, even if the Brucella blood titers become negative. They suggest that treatment efficacy should be monitored through serum and joint‐aspirate Brucella titers. Monotherapy with a fluoroquinolone or trimethoprim‐sulfamethoxazole has an unacceptable high‐risk rate of relapse [70]. Complete eradication of the microorganism is

30

Bone and Joint Infections

difficult to achieve, and relapse does occur, especially when the disease is caused by B. melitensis. The most frequent cause of relapse includes failure to complete treatment and unrecognized localized foci of infection [66]. The relapse is confirmed by the isolation of brucellae from blood or other tissues of a patient with recurrent symptoms. TSE was performed in 13/35 cases. The optimal time to reimplantation is unknown, because there is no consistent test to ensure successful eradication. Generally, patients with loose prostheses were treated with a TSE with a long interval (6 weeks–6 months between the stages). Prolonged antimicrobial therapy without resection of the prosthesis was chosen in cases without implant loosening (9/35), and yielded some success. DAIR and OSE or resection arthroplasty alone were also reported. Treatment was administered until resolution of infection and sterilization of synovial fluid cultures. Failure of antimicrobial therapy was documented in one case of B. abortus THA infection [71]. The patient was subsequently treated with OSE followed by one year of combination antimicrobial therapy until negative serologic titers were reached. The duration of follow‐up for all reported cases varied from six months to two years.

Francisella tularensis PJI caused by F. tularensis was recently described [72–74]. Cooper et al. [73] reported a case of a chronic TKA PJI in a 68‐year‐old man with rheumatoid arthritis treated with methotrexate who had a history of wood tick bite. He presented six months after TKA with persistent, low‐grade infection. It was presumed that asymptomatic lympho‐hematogenous seeding of F. tularensis was the likely mechanism of infection. Another two cases were reported by Chrdle et al. [72] in immunocompetent individuals in the late postoperative period diagnosed initially with culture‐negative PJIs. Their risk factors for tularemia were exposure to a rabbit barn and outdoor gardening in a tularemic area, respectively. A fourth case involving a THA was recorded in a hunter with no known recent exposures to tularemia [74]. Two of the reported cases had a lesion in the lower extremity, suggestive of a tick bite. Microbiologic Diagnosis F. tularensis is suspected on the basis of growth, morphology, and a weak catalase test. The microorganism is fastidious. Most strains require cysteine or cystine for growth. Prolonged incubation time to 14 days was recommended for culture negative PJIs [72]. Species identification can be accomplished by antisera, biochemical testing, and 16S rRNA sequencing [75]. Antimicrobial susceptibilities have not changed in years, and doxycycline, aminoglycosides, and fluoroquinolones alone or in combination therapy are the mainstay of therapy. Rifampin has a relatively low MIC against F. tularensis, but clinical experience is lacking. Therapy Antibiotic treatment with either ciprofloxacin/rifampin, doxycycline/gentamicin, or ­doxycycline alone was given between 3 weeks and 12 months. TSE, aspiration alone with antibiotics and DAIR without chronic oral suppression were surgical modalities ­performed in the reported cases. Repeated courses of antibiotics were needed in some of the reported cases.

3  Unusual Microorganisms in Periprosthetic Joint Infection

31

Yersinia enterocolitica To date only four cases of PJI due to Y. enterocolitica have been reported in the literature [76,77]. This infection is primarily associated with gastrointestinal symptoms (diarrhea, abdominal pain and mesenteric lymphadenitis in children, and fever). Severe symptoms such as acute ileitis, myocarditis, and septicemia are more commonly reported in adults. It may also cause reactive arthritis. Pigs and cattle are the primary reservoir of Y. enterocolitica. Reported cases occurred in elderly patients and involved a TKA or THA several years after implantation. The presentation of infection was rather acute with high fever and an inflamed prosthetic joint. Only one patient with diabetes mellitus who took iron supplements for anemia had diarrhea prior to the onset of symptoms. In this case, it is likely that the microorganism seeded to the TKA following a transient bacteremia after the initial enteritis Yersinia enterocolitica bacteremia without apparent source was documented in another case [77]. Iron supply associated with hemosiderin deposition due to hemarthrosis may be the main risk factor, at least in one case [76]. Microbiologic Diagnosis Y. enterocolitica grows well on most enteric media with the exception of Salmonella‐ Shigella agar. Serology testing by using bacterial agglutination may produce spurious results, especially in individuals infected with serogroups other than O:8. Geographic location and cross‐reactivity with other members of Enterobacteriacceae family, Brucella and Rickettsia spp. are also limitations to the agglutination method. ELISA technique has a higher sensitivity in assessing antibodies (IgG and IgA) not detected by bacterial agglutination. However, in the absence of culture‐proven Y. enterocolitica infection, the significance of a single positive antibody test is difficult to assess. In the reported cases, culture from joint fluid, operative specimens, and blood, as well as serology, identified Y. enterocolitica, serotype 0:3 and 0:9, respectively. All serogroups are susceptible to imipenem and aztreonam. Fluoroquinolones and broad‐spectrum cephalosporins, often in combination with an aminoglycoside, have resulted in a successful outcome in patients with extraintestinal Y. enterocolitica infection. The pattern of susceptibility of the four major serogroups (O:3, O:5,27, O:8, and O:9) to ampicillin, carbenicillin, cephalothin, cephaloridine, cephalexin, and cefoxitin was distinct to each serogroup. Therapy Removal of the prosthesis has been necessary for full recovery. However, a 90‐year old patient was treated with DAIR and chronic oral suppression with ciprofloxacin, with an initial very slow response [78]. Another case was cured with oral ciprofloxacin for six weeks, but the duration of the follow‐up was not reported [76]. The optimal duration of treatment is unknown.

Pasteurella multocida P. multocida infection typically localizes in skin and soft tissues. P. multocida, a small Gram‐negative organism, is part of the normal mouth flora of many animals. Thirty‐two cases of PJI due to P. multocida have been reported [79,80]. Most cases involve a TKA, and many patients were immunocompromised patients (diabetes mellitus, rheumatoid

32

Bone and Joint Infections

arthritis, acute leukemia, breast cancer). Almost all patients presented with a history of animal bite or animal contact or lick to the lower extremity containing the prosthetic joint. Two TKA PJIs were described after cat bites to the upper extremity [81,82]. This might reflect local propagation from the portal of entry via damaged lymphatic vessels or might be secondary to hematogenous seeding [81–83]. Cases of Pasteurella PJI transmitted by dog lick [84] or close contact with a dog [85] have been described. The majority of P. multocida PJI occurred in women. The reason for gender predilection could be that women have a higher risk of exposition as primary caretakers of pets. P. multocida PJI is usually an acute illness. Microbiologic Diagnosis Pasteurella can be cultured on blood or chocolate agar, preferably with an atmosphere containing 5% CO2. Some commercial biochemical identification systems are able to identify strains of P. multocida. Although Pasteurella is usually susceptible to penicillin, empiric treatment with a penicillinase‐resistant penicillin may be warranted pending in vitro susceptibility results. This is due to the fact that some isolates are β‐lactamase producers [86]. Therapy Cases of P. multocida PJI were successfully treated with TSE [85,86], OSE [87,88], DAIR [79], or resection arthroplasty [79,81].The duration of antimicrobial therapy in these cases was variable, ranging from 3 to 10 weeks of parenteral therapy followed in some cases by 2–3 weeks of oral antimicrobial therapy [81, 88]. Success was also reported in cases where the infected prosthesis was retained using surgical debridement [89,90], aspiration [91], or antimicrobial therapy alone [92,93]. On the other hand, several case reports of treatment failure occurred in patients treated with antimicrobial therapy with or without joint aspiration [82,84]. Guion and Sculco [86] suggest that optimal surgical therapy for these infections should include debridement of the involved tissues and removal of all foreign material. However, DAIR combined with prolonged antimicrobials may be sufficient for the treatment of P multocida PJIs [79]. Some authors suggest that following cat or dog bites especially immunocompromised patients with a total joint arthroplasty should be instructed to take a penicillin prophylaxis, in order to prevent hematogenous seeding on the implant [79,90].

Coxiella burnetii Coxiella burnetii, which causes Q‐fever, is an organism that would traditionally be considered in culture‐negative PJIs. Eight cases have been described in the literature, five involving a THA and three involving a TKA [94,95]. Only one case had a documented contact with sheep [96]. Other cases had either no risk factors or local steroid soft‐tissue injections or history of IVDU. Local signs were the main presenting symptoms, and one patient was asymptomatic, Q fever being diagnosed at the time of revision THA [96]. In six cases, the diagnosis was established by specific PCR. A serological profile with strongly positive phase 1 IgG serology also supported the diagnosis in all cases. It is unclear whether C burnetii is capable of biofilm formation [95]. In ostearticular infections, a

3  Unusual Microorganisms in Periprosthetic Joint Infection

33

treatment of 18 months with hydroxychloroquine and doxycycline has been suggested as the best option [97]. Alternative treatments include ciprofloxacin or trimethoprim‐­ sulfamethoxazole in combination with doxycycline. Most of the PJI cases had a good outcome. However, one case required resection arthroplasty and transfemoral amputation [98], and two others required additional surgeries [96,99].

Campylobacter Campylobacter spp. are major causes of gastrointestinal infections in humans. Septic native arthritis is uncommon, and usually affects elderly patients, immunocompromised patients, and patients with alcoholism, cirrhosis, chronic lymphatic leukemia, cancer, or diabetes mellitus. Infections caused by C. jejuni, C. coli, and C. lari can be acquired via contact with farm animals, chickens, seagulls, and nonhuman primates. A fatal case of C. lari THA‐PJI with bacteremia has been reported in an 81‐year‐old immunocompetent patient who suffered from a non‐specific infectious syndrome without diarrhea for 3 weeks [100]. Not all cases have a preceding diarrheic episode. Bacteremia is often preceded or accompanied by diarrhea [101]. Cases of THA PJI caused by C. jejuni [102, 103], lari [100], and fetus [104] have been reported. Microbiologic Diagnosis Campylobacter spp. are fastidious, slowly growing Gram‐negative bacilli which require selective, enriched media with prolonged incubation under microaerophilic conditions. Generally, Campylobacter spp. are not easily visualized with the safranin counterstain commonly used in Gram staining. Therefore, its identification is difficult, especially from a periprosthetic joint site, at which Campylobacter is infrequently or unexpectedly isolated. Molecular testing is useful in identification of campylobacters, especially in distinguishing C. lari from C. jejuni, since they both share common phenotypic properties [100]. A previously described C. jejuni/C. coli real time PCR assay targeting cadF and designed for testing stool was used for diagnosis of C. jejuni in the synovial fluid [105]. Determination of in vitro antimicrobial susceptibility remains controversial, since most infections are self‐limited, but it should be performed in extraintestinal or severe cases. C. jejuni is susceptible to macrolides, fluoroquinolones, aminoglycosides, tetracyclines, and chloramphenicol. C. jejuni is resistant to trimethoprim and most β‐lactam antibiotics [106]. Of concern is the increased incidence of resistance to antimicrobials in campylobacters, particularly to fluoroquinolones [106–108]. Therapy In general, surgical treatment of Campylobacter PJI is dictated by the clinical presentation, underlying conditions, and the status of the prosthetic device. Some cases were treated with DAIR, followed by chronic oral suppression. One patient with C. lari bacteremia died of septic complications [100]. Other patients were managed with antimicrobial therapy without surgery. One of them received one month of chloramphenicol and died after two months of an unrelated cardiac cause [109]. The second patient received eight weeks of parenteral antimicrobials (main treatment erythromycin and ciprofloxacin), and had a good outcome at six months of follow‐up [102].

34

Bone and Joint Infections

Erysipelotrix rhusiopathiae E. rhusiopathiae is the causative organism of erysipeloid. It is associated with exposure to domestic swine, muskox die‐offs, and linked to the emergence of a new disease syndrome in Arctic fox. Several cases of TKA and THA PJI have been reported [110–114]. All but one case had exposure to animals, owned a hunting dog, or was a butcher. Two patients were immunocompromised due to steroid use, rheumatoid arthritis, lupus nephritis, or chronic alcoholism. Systemic symptoms are generally absent. Diagnosis relies on isolation of microorganism in culture. Molecular diagnosis includes 16S rRNA sequencing. PCR on joint fluid may be a sensitive and specific way to detect the microorganism, particularly in cases with high index of suspicion but negative cultures [111]. E. rhusiopathiae is intrinsically resistant to vancomycin and aminoglycosides. It is susceptible to penicillins, broad spectrum cephalosporins, and fluorquinolones [115]. Treatment requires a combination of surgical intervention and antimicrobial therapy. The surgical treatment involved resection arthroplasty, TSE in the reported cases.

Anaerobic Microorganisms Clostridium difficile C. difficile rarely causes extra‐colonic disease and it is an unusual cause of bone and joint infection. Clostridium PJI typically occurs in patients with underlying gastrointestinal disease. C. difficile has been identified in hip, knee, and shoulder PJIs [116–118]. Previous antimicrobial use was reported in all cases of C. difficile THA. Two cases presented as late chronic infections and occurred after C. difficile‐associated diarrhea (CDAD) at variable intervals. Mc Carthy et al. [119] reported a case of C. difficile THA PJI that occurred 12 months after resolution of CDAD. The strain isolated from the THA was identical to the one isolated from stool 12 months earlier using pulse‐field gel electrophoresis. Therapy All reported cases were treated with metronidazole. Resection arthroplasty and limb amputation after failure of open arthrotomy and antimicrobial therapy were required in two patients, respectively. Death occurred in one case as a result of complications of primary disease.

Actinomyces Actinomyces spp. rarely cause PJIs. Comorbid conditions included obesity, IV‐drug use, diabetes mellitus, dental procedures, and IUD [120–129]. Patients presented with late onset symptoms following primary TJA or revision surgery, the interval ranging from 20 days and 11 years. Both monomicrobial and polymicrobial infections have been reported. Microbiologic Diagnosis Actinomyces spp. is difficult to culture, and in the lab, growth can take 5–20 days [130]. 16S rRNA gene sequencing may help with detection of Actinomyces. The Matrix‐Assisted Laser Desorption Ionization‐Time of Flight Mass Spectrometry (MALDI‐TOF) can provide rapid and accurate identification [124].

3  Unusual Microorganisms in Periprosthetic Joint Infection

35

Therapy Among reported cases, most patients were treated with TSE, but resection arthroplasty, DAIR [121], or OSE have been reported [128]. Following reimplantation, some authors have suggested a six‐week to three‐month course of amoxicillin or oral penicillin [124]. Actinomyces spp. are susceptible to beta‐lactams, but oxacillin, cloxacillin, and cephalexin are not effective. Doxycycline has been used in cases of penicillin allergy. Reported cases of Actinomyces PJIs were treated for six to nine weeks with oral or parenteral beta‐lactams, followed by four weeks to one year, or even indefinite, oral suppressive therapy [120,121,123].

Other Anaerobes Cases of PJI caused by Veionella dispar, V. parvula, Prevotella melaninogenica, and Clostridium perfringens were reported or included in PJI series [131–134]. Clostridium septicum PJI has been associated with intestinal malignancy [135,136].

Mycobacteria Mycobacterium tuberculosis Infection of the musculoskeletal system represents 1–5% of tuberculosis cases [137–139]. Few case series of M. tuberculosis total joint arthroplasty infections have been reported. In a retrospective study on 2116 episodes of PJI over a 22‐year period, only 7 (0.3%) were due to M. tuberculosis [139,140]. Tuberculous PJI usually involves hip, knee, or shoulder, and can result most commonly from either local reactivation or occasionally from ­hematogenous spread. M. tuberculosis TJA infection in patients without prior history of tuberculosis has been reported [139,141,142]. The risk of reactivation of M. tuberculosis in patients undergoing THA or TKA for quiescent tuberculous native septic arthritis varies between 0–31%. It is higher for patients receiving a TKA (27%) than for those receiving a THA (6%) [140–142]. The use of antituberculous therapy at the time of arthroplasty for latent M. tuberculosis septic arthritis may be reasonable for patients who have not received modern antimycobacterial chemotherapy. Alternatively, preoperative isoniazid prophylaxis for their previous M. tuberculosis septic arthritis could be considered [141–145]. The duration of prophylaxis remains unknown. In patients with an underlying immunodeficiency, history of tuberculous infection, or with tuberculosis risk factors postoperative tuberculosis prophylaxis may be advisable [146]. Obtaining mycobacterial cultures at the time of arthroplasty should be done in these cases. PPD skin testing should be performed prior to total joint arthroplasty in patients from a high prevalence area, in those who have a history of native joint septic arthritis due to an unknown pathogen, or when the underlying joint disease is unknown. The majority of patients with M. tuberculosis PJI are PPD positive, but a negative test has been reported by Tokumoto et  al. [147]. Quantiferon‐B Gold, can be helpful, but it cannot distinguish between latent and active tuberculosis. The clinical course of tuberculous PJI can present in two patterns. First, patients are recognized at the time of arthroplasty based on histologic or microbiologic evidence of M. tuberculosis infection; second, tuberculosis is only recognized in the late postoperative period (>6 weeks). In the latter situation, M. tuberculosis PJI often presents insidiously, over weeks to months. A draining sinus is commonly seen and was present in all cases described by Berbari et al. [140].

36

Bone and Joint Infections

Microbiologic Diagnosis Currently, the gold standard for diagnosis includes a joint fluid or synovial fluid analysis for acid‐fast bacilli culture and histopathology [146]. However, negative findings on histopathological specimens or cultures do not necessarily prove the absence of tuberculous infection [143,146]. PCR testing for extra pulmonary specimens has a sensitivity between 53.7% and 100%, depending on sample selection [146]. Neogi et al. [139] reported on a 73‐year‐old female with tuberculous PJI 14 years following a TKA who had negative synovial fluid and joint fluid cultures, but had a positive synovial tissue M. tuberculosis PCR. Therapy In some cases, a coinfecting bacterial pathogen was reported [139,140]. In vitro susceptibility testing should be performed for all M. tuberculosis isolates, because of the emergence of resistance. Initial therapy should include isoniazid, rifampin, and pyrazinamide, with the addition of ethambutol or streptomycin in case of suspected isoniazid resistance [139,148] (see Chapter 19). The optimal medical and surgical therapy for M. tuberculosis PJI is unknown. Prolonged antituberculous regimen including rifampin for a median duration of 12 months in the literature, 14 months in a French multicenter study, can be curative, even in the absence of surgery, as the capacity of M. tuberculosis for adherence and biofilm formation is much lower as compared to that of staphylococci [149]. For patients with late onset M. tuberculosis PJI, medical treatment alone is usually unsuccessful, and removal of the prosthesis is often required. TSE [140,149–151], partial OSE [140,149– 153], DAIR [140,141,147,149,154,155], and medical management alone have all been performed. Failure occurred in 50% of the patients treated with DAIR, indicating that exchange of the device should be favored [140,141,147,149,154,155].

Mycobacterium bovis Cases of Mycobacterium bovis prosthetic joint infection have been reported, mainly as complication of Bacillus‐Calmette‐Guerin instillation in patients with bladder cancer [156-163].

Non‐tuberculous mycobacteria Mycobacterium chelonae and fortuitum have been rarely described as a cause of PJI. These rapid growing mycobacteria are non‐pigmented mycobacteria found usually in soil or water. Several articles reported M. fortuitum PJI involving both hips and knees [157– 169], but very few described M. chelonae PJI [165,170,171] and M. abscessus PJI [172– 179]. Most of the PJIs caused by M. fortuitum occurred in the early postoperative period. M. chelonae usually manifests as late infection. The presentation is usually acute, with drainage, abscess, and fistulae formation. Microbiologic Diagnosis Mycobacterium chelonae grows more slowly than common bacteria but more rapidly than other mycobacteria on agar plates (5–7 days). Often cultures of M. chelonae are

3  Unusual Microorganisms in Periprosthetic Joint Infection

37

reported as negative, or can be misidentified as a coryneform. M. chelonae could also be misidentified as Nocardia asteroides. Primary cultures should be incubated for six weeks because original growth may not be observed until several weeks of incubation. Differentiation of M chelonae from M. fortuitum by biochemical means is important because of their different drug susceptibilities. Susceptibility testing by disk method has a definite time advantage over the iron uptake test and appears to be equally accurate [179,180]. The recommended susceptibility testing method for rapid growing mycobacteria is the broth microdilution technique, with MIC determinations and resistance breakpoints similar to those used for other bacterial species (NCCLS). Therapy M. fortuitum is much more drug susceptible than M. chelonae ssp. abscessus. Essentially all isolates have in vitro susceptibility to achievable serum levels of amikacin, cefoxitin, imipenem, sulfonamides, fluoroquinolones [179,181]. M. fortuitum isolates may develop resistance in spite of multidrug regimens, therefore repeating susceptibility testing during treatment is important when failure is suspected [179,182]. M. abscessus is a very difficult microorganism to control, because it is highly resistant to antibiotics. It is also resistant to common disinfectants [183]. It is usually susceptible to amikacin and cefoxitin [181,183], and some isolates are moderately susceptible to ­imipenem. Unlike M. fortuitum, isolates of M. chelonae ssp. abscessus are usually resistant in vitro to most of the potentially useful oral antimicrobial agents (doxycycline, erythromycin, and sulfisoxazole) [183–186]. Linezolid and tedizolid seem to be promising agents against M.  fortuitum, chelonae, and some M. abscessus isolates. Proposed MIC breakpoints for linezolid are ≤8 μg/ml for susceptible strains, 16 μg/ml for moderately susceptible, and 32 μg/ ml, respectively, for resistant strains [183,187,188]. Linezolid‐resistant strains of M. abscessus have been reported [177]. Tigecycline may also be useful for treatment of M abscessus PJIs. Clofazimine is a second line alternative for these infections. Antimicrobial therapy alone or in combination with DAIR has been proved to be ineffective in PJI caused by M fortuitum. Most patients required removal of the prosthesis or arthrodesis. Prolonged parenteral antimicrobial therapy for six weeks, followed by three to six months of oral therapy with bacteriologic evidence of complete elimination of the infection before reimplantation, has been suggested [167,177]. M. chelonae TKA infection required six weeks of cefoxitin and amikacin, followed by administration of trimethoprim‐sulfamethoxazole for a total of three months of therapy. OSE and chronic ciprofloxacin suppression had a good outcome after two‐year follow‐up [170,177]. Removal of the implant is essential in M. abscessus PJI as this microorganism can display a smooth forming phenotype that can be a significant impediment in achieving a microbiological cure [189]. Cases of M. abscessus were treated with resection arthroplasty [172,173,179,189], DAIR [179], TSE [175,183], or arthrodesis [183]. Antimicrobial regimens are complex and directed by the susceptibility data. The recommended duration of therapy is 6–12 months. A TSE combined with negative aspiration cultures following at least two months of antimicrobial holiday was recommended before reimplantation for M abscessus PJI [175]. The role of amikacin‐impregnated cement spacer for M. abscessus PJI is debatable. Although the incorporation of antibiotic cement has emerged as the standard of care in the treatment of PJIs, the efficacy of this adjunct has not been corroborated by clinical trials. In our experience, the outcomes of M. abscessus PJI have been poor, though a successful outcome has been reported [172,175,177,183].

38

Bone and Joint Infections

Mycobacterium avium complex Mycobacterium avian complex (MAC) is rarely associated with PJIs. The infection tends to occur in immunocompromised hosts (HIV, transplantation, SLE, rheumatoid arthritis) [190–194]. In contrast to M. tuberculosis PJI, Mycobacterium avium complex PJI occurs as a result of recent hematogenous dissemination, as opposed to reactivation. McLaughlin et al. [192] described a case of M. avium complex THA infection in a 20‐ year‐old man with AIDS who also had M. avium complex bacteremia and histopathological evidence of disseminated M. avium complex infection in both of his THAs. Microbiologic Diagnosis Mycobacterium avium complex can be a contaminant which complicates the interpretation of the culture result [195]. On pathology, granulomas are poorly formed and host inflammatory response is minimal [196]. Therapy Because of the rarity of MAC PJI, optimal management is not clearly defined. Susceptibilities should be obtained to guide the treatment. A three‐drug regimen of a macrolide, ethambutol, and a rifamycin is recommended for 6–12 months, in addition to resection arthroplasty with debridement [193,197]. Resection arthroplasty followed by appropriate antimycobacterial therapy provides the best outcome. If DAIR is performed, chronic oral suppression may needed to prevent relapse [173,191,193].

Other Microorganisms Mycoplasma M. hominis, M. pneumonia, and M. salivarium were reported as rare causes of PJI [198– 202]. M. hominis PJI should be suspected in patients with clinically infected joint with purulent aspirate, negative Gram stain, and negative standard cultures. Mycoplasma should be considered in the differential diagnosis in patients with culture‐negative PJIs, particularly in those with hypogammaglobulinemia. A case of Mycoplasma hominis of a prosthetic knee after transurethral biopsy in an immunocompetent patient was recently described [202]. The microbiology laboratory should be asked to look specifically for M. hominis. Gram stain of the joint aspirate is unrevealing. Serum and joint fluid antibody levels become detectable or rise during the course of the illness. 16S rRNA sequencing identified Mycoplasma hominis in a prosthetic knee [202]. Metagenomic shotgun sequencing detected Mycoplasma salivarium in a patient with TKA PJI and hypogammmaglobulinemia [199]. Methods available for susceptibility testing of M. hominis are not standardized, and do not correlate with the clinical outcome. Many strains are susceptible in vitro to tetracyclines, clindamycin and are moderately susceptible to rifampin. M. hominis is generally resistant to aminoglycosides, beta‐lactam antibiotics, vancomycin, sulfonamides, trimethoprim, and erythromycin. Fluoroquinolones are usually active in vitro against M. hominis, but resistance can be induced in vitro by exposing the microorganism to increasing concentrations of fluoroquinolones. Limited information suggests that M. hominis is susceptible in vitro to

3  Unusual Microorganisms in Periprosthetic Joint Infection

39

linezolid and quinupristin‐dalfopristin although no clinical trials have been undertaken using these agents in M. hominis infections [203].

Ureplasma U. urealyticum, parvum have been recently described as a rare cause of THA and TKA PJI and diagnosed by 16S rRNA sequencing [204–206].

Echinococcus Echinococcal THA PJI was reported in three patients [207–209]. Serologic tests may help confirm a suspected case of echinococcosis. A negative serologic test does not rule out echinococcosis. The Casoni intradermal test and the Weinberg complement fixation test confirmed the diagnosis in one reported case [207]. Pathology made the diagnosis in the second case [208]. The management of bone infestation by Echinococcus requires complete resection of the involved area. Adjunctive chemotherapy before and after surgery appears to reduce the risk of recurrence by inactivating protoscolices and lessening the tension of the cysts for easier cyst removal [210]. Removal of the prosthesis and prolonged treatment with albendazole is required in cases when complete resection of the cysts is not technically feasible.

Tropheryma whipplei PJI caused by T. whipplei has been described in a patient with a TKA two years after complete cure of Whipple’s disease in one patient, and in a patient with THA with previously undiagnosed Whipple’s disease [211,212]. Microbiologic diagnosis is usually made by PCR analysis of the joint fluid. A positive PCR result for small bowel tissue in the presence of undiagnosed joint infection may support a diagnosis of T. whipplei joint infection. Attempts to culture the causative organism are unsuccessful. DAIR of the knee prosthesis and chronic oral antimicrobial suppression with trimethoprim‐sulfamethoxazole followed by pristinamycin controlled the infection after one year of follow‐up [211]. The patient with THA PJI was treated with OSE followed by nine months of trimethoprim sulfamethoxazole with good short‐term outcome [212].

Borrelia burgdorferi TKA PJI caused by B. burgdorferi has been described in four patients [213–215]. All cases occurred in an endemic area. A history of tick exposure was not always apparent. The onset of symptoms was acute in all presented cases. The diagnosis for these culture ­negative PJI was made by Lyme serology and Lyme PCR in the synovial fluid. Optimal therapy remains unknown, but four to six weeks of antimicrobial therapy (ceftriaxone, doxycycline for the reported cases) seemed to be adequate. Surgical intervention entailed DAIR in two patients, TSE and no surgery, respectively, in two other patients [213–215]. In regions with high prevalence of Lyme disease, synovial fluid Lyme PCR testing in conjunction with Lyme antibody testing should be done, particularly in cases of culture‐ negative PJI.

40

Bone and Joint Infections

Key Points ●●

●●

●●

●●

When treating patients with PJI, an appropriate exposure history is of great clinical interest, because adequate antimicrobial therapy depends on the correct etiologic ­diagnosis. This is especially important in situations, when routine bacterial cultures fail to identify a microorganism despite proof of PJI with non‐microbiological criteria. In patients with culture‐negative PJI, it is crucial to maintain a high index of suspicion, because unusual or fastidious microorganisms require special stains and culture conditions, or modern molecular methods (see Chapter 4). Communication between the microbiologist and the orthopedic infectious disease specialist is extremely important for final identification of these microorganisms. Surgical management of PJI‐patients caused by rare microorganisms cannot be standardized, because no large series with defined treatment are published.

References 1. Askar M, Bloch B, Bayston R. Small‐colony variant of Staphylococcus lugdunensis in prosthetic joint infection. Arthroplast Today. 2018;4(3):257–260. 2. Brandt CM, Duffy MC, Berbari EF, et al. Staphylococcus aureus prosthetic joint infection treated with prosthesis removal and delayed reimplantation arthroplasty. Mayo Clin Proc. 1999;74(6):553–558. 3. Sampathkumar P, Osmon DR, Cockerill FR, 3rd. Prosthetic joint infection due to Staphylococcus lugdunensis. Mayo Clin Proc. 2000;75(5):511–512. 4. Weightman NC, Allerton KE, France J. Bone and prosthetic joint infection with Staphylococcus lugdunensis. J Infect. 2000;40(1):98–99. 5. Nagy MT, Hla SM, Keys GW. Late Streptococcus bovis infection of total knee replacement complicated by infective endocarditis and associated with colonic ulcers. BMJ Case Rep. 2013;2013. 6. Thompson JC, Goldman AH, Tande AJ, et  al. Streptococcus bovis Hip and Knee Periprosthetic Joint Infections: A Series of 9 Cases. J Bone Jt Infect. 2020;5(1):1–6. 7. Wilde AH, Ruth JT. Two‐stage reimplantation in infected total knee arthroplasty. Clin Orthop. 1988(236):23–35. 8. Emerton ME, Crook DW, Cooke PH. Streptococcus bovis‐infected total hip arthroplasty. J Arthroplasty. 1995;10(4):554–555. 9. Clarridge JE, 3rd, Attorri SM, Zhang Q, et  al. 16S ribosomal DNA sequence analysis distinguishes biotypes of Streptococcus bovis: Streptococcus bovis Biotype II/2 is a separate genospecies and the predominant clinical isolate in adult males. J Clin Microbiol. 2001;39(4):1549–1552. 10. Eggelmeijer F, Petit P, Dijkmans BA. Total knee arthroplasty infection due to Gemella haemolysans. Br J Rheumatol. 1992;31(1):67–69. 11. von Essen R, Ikavalko M, Forsblom B. Isolation of Gemella morbillorum from joint fluid. Lancet. 1993;342(8864):177–178. 12. Leung DT, Davis EM, Quian Q et al. First report of prosthetic joint infection by Gemella sanguinis and associated “pseudosatelliting” phenomenon on culture. J Clin Microbiol. 2011;49(9):3395–3397. 13. von Essen R. Blood culture bottles improve microbe isolation rate from joint fluids. Am J Clin Pathol. 1989;91(4):501–503. 14. Nieman RE, Lorber B. Listeriosis in adults: a changing pattern. Report of eight cases and review of the literature, 1968‐1978. Rev Infect Dis. 1980;2(2):207–227. 15. Bush LM, Abdulah A, Perez MT. Listeria monocytogenes prosthetic joint infections. A review a propos a case report. 2014; Infect Dis. Clin Pract: 1–4.

3  Unusual Microorganisms in Periprosthetic Joint Infection

41

16. Allerberger F, Kasten MJ, Cockerill FR, 3rd, et  al. Listeria monocytogenes infection in ­prosthetic joints. Int Orthop. 1992;16(3):237–239. 17. Diaz‐Dilernia F, Costantini J, Nicolino TI et al. Unusual Listeria monocytogenes hematogenous infection in total knee replacement treated with one‐stage revision surgery. Arthroplasty today 2019:296–300. 18. Seo Y, Noh YS, Wie SH, et al. Prosthetic knee joint infection due to Listeria monocytogenes bacteremia in a diabetic female. Korean J Intern Med. 2016 31:616–619. 19. Tabib W, Guiffault P, Lemort CB, et al. Prosthetic hip joint infection caused by Listeria monocytogenes. Acta Orthop Belg. 2002;68(2):182–186. 20. Arnal C, Man H, Delisle F, M’Bappe P, et al. Nocardia infection of a joint prosthesis complicating systemic lupus erythematosus. Lupus. 2000;9(4):304–306. 21. Robinson D, Halperin N. Nocardia asteroides infection of an Austin‐Moore hemiarthroplasty in a nonimmunocompromised host A case report. Bull Hosp Jt Dis Orthop Inst. 1989; 49(1):107–110. 22. Hadeed MM, MacDonell JR, Dempsey IJ et al. Chronic Nocardia cyriacigeorgica Periprosthetic Knee Infection Successfully Treated with a Two‐Stage Revision: A Case Report. JBJS Case Connect. 2017;7(4):e74. 23. Laurent F, Rodriguez‐Villalobos H, Cornu O, et al. Nocardia prosthetic knee infection successfully treated by one‐stage exchange: case report and review. Acta Clin Belg. 2015;70(4): 287–290. 24. Palmer DL, Harvey RL, Wheeler JK. Diagnostic and therapeutic considerations in Nocardia asteroides infection. Medicine. 1974;53(5):391–401. 25. Ambaye A, Kohner PC, Wollan PC, et al. Comparison of agar dilution, broth microdilution, disk diffusion, E‐test, and BACTEC radiometric methods for antimicrobial susceptibility testing of clinical isolates of the Nocardia asteroides complex. J Clin Microbiol. 1997;35(4): 847–852. 26. Wallace RJ, Jr, Septimus EJ, Williams TW, et al. Use of trimethoprim‐sulfamethoxazole for treatment of infections due to Nocardia. Rev Infect Dis. 1982;4(2):315–325. 27. Pidoux O, Argenson JN, Jacomo V, et al. Molecular identification of a Dietzia maris hip prosthesis infection isolate. J Clin Microbiol. 2001;39(7):2634–2636. 28. Larkin JA, Lit L, Sinnott J, et  al. Infection of a knee prosthesis with Tsukamurella species. South Med J. 1999;92(8):831–832. 29. Harrington RD, Lewis CG, Aslanzadeh J, et al. Oerskovia xanthineolytica infection of a prosthetic joint: case report and review. J Clin Microbiol. 1996;34(7):1821–1824. 30. Funke G, von Graevenitz A, Clarridge JE, 3rd et al. Clinical microbiology of coryneform bacteria. Clin Microbiol Rev. 1997;10(1):125–159. 31. Reboli AC, Bryan CS, Farrar WE. Bacteremia and infection of a hip prosthesis caused by Bacillus alvei. J Clin Microbiol. 1989;27(6):1395–1396. 32. Ha J, Park YJ, Kim YJ et al. Late prosthetic joint infection and bacteremia by Bacillus cereus confirmed by 16S rRNA sequencing and hip joint tissue pathology. Ann Clin Microbiol 2016;19(2):54–57. 33. Rae S, Webley M, Snaith ML. Salmonella typhimurium arthritis in rheumatoid disease. Rheumatol Rehabil. 1977;16(3):150–151. 34. Widmer AF, Colombo VE, Gachter A, et  al. Salmonella infection in total hip replacement: tests to predict the outcome of antimicrobial therapy. Scand J Infect Dis. 1990;22(5): 611–618. 35. Samra Y, Shaked Y, Maier MK. Nontyphoid salmonellosis in patients with total hip replacement: report of four cases and review of the literature. Rev Infect Dis. 1986;8(6):978–983. 36. Arda B, Sipahi OR, Yamazhan T, et al. Salmonella enteritidis related prosthetic joint infection. West Indian Med J. 2006;55(6):454–455. 37. Ekinci M, Bayram S, Akgul T. et al. Periprosthetic joint infection caused by Salmonella‐ Case reports of two azathioprine and prednisolone induced immunocompromised patients. Hip Pelvis 2017;29(2):139–144.

42

Bone and Joint Infections

38. Gupta A, Berbari EF, Osmon DR et al. Prosthetic joint infection due to Salmonella species: a case series. BMC Infect Dis. 2014;14:633. 39. Lo I‐F, Chang. H‐C. Salmonella Septic Arthritis in A Patient with A Hip Implant: A Case Report. Intern J Gerontology 2018;12(4):344–347. 40. Tsukayama DT, Estrada R, Gustilo RB. Infection after total hip arthroplasty. A study of the treatment of one hundred and six infections. J Bone Joint Surg Am. 1996;78(4):512–523. 41. Hashimoto Y, Itho Y, Fujinaga Y, et  al. Development of nested PCR based on the ViaB sequence to detect Salmonella typhi. J Clin Microbiol. 1995;33(3):775–777. 42. Glynn MK, Bopp C, Dewitt W, et al. Emergence of multidrug‐resistant Salmonella enterica serotype typhimurium DT104 infections in the United States. N Engl J Med. 1998;338(19):1333–1338. 43. Chiu CH, Wu TL, Su LH, et al. The emergence in Taiwan of fluoroquinolone resistance in Salmonella enterica serotype choleraesuis. N Engl J Med. 2002;346(6):413–419. 44. Day LJ, Qayyum QJ, Kauffman CA. Salmonella prosthetic joint septic arthritis. Clin Microbiol Infect. 2002;8(7):427–430. 45. Chen CM, Lu TC, Lo WH. Salmonella infection in total hip replacement‐‐report of successful reimplantation and review of the literature. Chung Hua i Hsueh Tsa Chih ‐ Chinese Med J. 1999;62(7):472–476. 46. Widmer AF, Colombo VE, Gächter A, et al. Salmonella Infection in Total Hip Replacement: Tests to Predict the Outcome of Antimicrobial Therapy. Scand J Infect Dis. 1990;22(5):611–618. 47. Fu TS, Ueng SW. Two‐staged revision total hip arthroplasty due to Salmonella infection: case report. Chang Gung Medical Journal. 2001;24(3):202–207. 48. Tattevin P, Cremieux A, Joly‐Guillou M. First case of Salmonella hirschfeldii infection of a prosthetic hip. Clin Microbiol Infect. 1998;4:228–230. 49. Madan S, Abbas D, Jowett RL, et al. Salmonella enteritidis infection in total knee replacement. Rheumatology (Oxford). 2001;40(1):112–113. 50. Rae S, Webley M, Snaith ML. Salmonella typhymurium arthritis in rheumatoid disease. Rheumatol Rehabil. 1977; 16:150–151. 51. Samra Y, Shaked Y, Maier MK. Nontyphoid Salmonellosis in Patients with Total Hip Replacement: Report of Four Cases and Review of the Literature. Rev Infect Dis 1986.;8(6):978–983. 52. Vikram HR, Buencamino RB, Aronin SI. Primary meningococcal arthritis in a prosthetic knee joint. J Infect. 2001;42(4):279–281. 53. Carral BB, Manoja EA, Cardenas SL. Neisseria meningitidis infecting a Prosthetic Knee Joint: A New Case of an Unusual Disease. Case Rep Infect Dis. 2017: 1–3. 54. McCarthy A, Broderick JM, AP. M. Neiserria meningitidis as a cause of septic arthritis: an unusual case of periprosthetic joint infection. Case Rep Infect Dis. 2020:1–3. 55. Söderquist B. Prosthetic hip joint infection caused by non‐capsulated Haemophilus influenzae. Scand J Infect Dis. 2014;46(9):665–668. 56. Khan S, Reedy S. Haemophilus influenzae infection is a prosthetic knee joint in a patient with CLL: a vaccine preventable disease. BMJ Case Rep. 2013:1–3. 57. Bezwada HP, Nazarian DG, Booth RE, Jr. Haemophilus influenzae infection complicating a total knee arthroplasty. Clin Orthop. 2002(402):202–205. 58. Borenstein DG, Simon GL. Hemophilus influenzae septic arthritis in adults. A report of four cases and a review of the literature. Medicine (Baltimore). 1986;65(3):191–201. 59. Jellicoe PA, Cohen A, Campbell P. Haemophilus parainfluenzae complicating total hip arthroplasty: a rapid failure. J Arthroplasty. 2002;17(1):114–116. 60. Manian FA. Prosthetic joint infection due to Haemophilus parainfluenzae after dental surgery. South Med J. 1991;84(6):807–808. 61. Ortega‐Andreu M, Rodriguez‐Merchan EC, Aguera‐Gavalda M. Brucellosis as a cause of septic loosening of total hip arthroplasty. J Arthroplasty. 2002;17(3):384–387.

3  Unusual Microorganisms in Periprosthetic Joint Infection

43

62. Rodriguez Zapata M, Gamo Herranz A, De La Morena Fernandez J. Comparative study of two regimens in the treatment of brucellosis. Chemioterapia. 1987;6(2 Suppl):360–362. 63. Weil Y, Mattan Y, Liebergall M, Rahav G. Brucella prosthetic joint infection: a report of 3 cases and a review of the literature. Clin Infect Dis. 2003;36(7):e81–86. 64. Agarwal S, Kadhi SK, Rooney RJ. Brucellosis complicating bilateral total knee arthroplasty. Clin Orthop. 1991(267):179–181. 65. Malizos KN, Makris CA, Soucacos PN. Total knee arthroplasties infected by Brucella melitensis: a case report. Am J Orthop. 1997;26(4):283–285. 66. Orti A, Roig P, Alcala R, Navarro V, et al. Brucellar prosthetic arthritis in a total knee replacement. Eur J Clin Microbiol Infect Dis. 1997;16(11):843–845. 67. Yagupsky P, Peled N, Press J, et al. Rapid detection of Brucella melitensis from blood cultures by a commercial system. Eur J Clin Microbiol Infect Dis. 1997;16(8):605–607. 68. Balkhair A, Al Maskari S, Ibrahim S et  al. Brucella Periprosthetic Joint Infection Involving Bilateral Knees with Negative Synovial Fluid Alpha‐Defensin. Case Rep Infect Dis. 2019(1–3). 69. Solís García del Pozo JI, Solera J. Systematic review and meta‐analysis of randomized clinical trials in the treatment of human brucellosis. Plos One 2012;7(2), e32090. 70. Lang R, Rubinstein E. Quinolones for the treatment of brucellosis. J Antimicrob Chemother. 1992;29(4):357–360. 71. Jones RB, Smith J, Hoffmann A, et al. Secondary infection of a total hip replacement with Brucella abortus. Orthopedics. 1983;6:184–186. 72. Chrdle A, Trnka T, Musil D et al. Francisella tularensis periprosthetic joint infections diagnosed with growth in cultures. J Clin Microbiol. 2019;57:1–7. 73. Cooper CL, Van Caeseele P, Canvin J, et al. Chronic prosthetic device infection with Francisella tularensis. Clin Infect Dis. 1999;29(6):1589–1591. 74. Rawal H, Patel A, Moran M. Unusual case of prosthetic joint infection caused by Francisella tularensis. UBMJ Case Rep. 2017:1–3. 75. Koneman EN, Allen SD, Janda WM. Other miscellanous fastidious Gram‐negative bacteria. Color atlas and textbook of diagnostic microbiology. 5th ed. Philadelphia: Lipincott‐Raven; 1997. p. 431. 76. Iglesias L, Garcia‐Arenzana JM, Valiente A, et al. Yersinia enterocolitica O:3 infection of a prosthetic knee joint related to recurrent hemarthrosis. Scand J Infect Dis. 2002;34(2): 132–133. 77. Oni JA, Kangesu T. Yersinia enterocolitica infection of a prosthetic knee joint. Br J Clin Pract. 1991;45(3):225. 78. Jalava‐Karvinen P, Oksi J, al. Rantakokko‐Jalava K et al. Yersinia enterocolitica infection of a prosthetic knee joint. Case report and review of the literature on dep sited infections caused by Yersinia enterocolitica. Advances Infect Dis. 2013 3:95–99. 79. Honnorat E, Seng P, Savini H et al. Prosthetic joint infection caused by Pasteurella multocida; a case series and review of the literature. BMC Infectious Dis. 2016;16(435):1–7. 80. Lam PW, Page A. Pasteurella multocida non‐native joint infection after dog lick. A case report describing a complicated two ‐stage revision and a comprehensive review of the literature. Can J Infect Dis Med Microbiol 2015;26(4):212–217. 81. Gabuzda GM, Barnett PR. Pasteurella infection in a total knee arthroplasty. Orthop Rev. 1992;21(5):601, 4–5. 82. Orton DW, Fulcher WH. Pasteurella multocida: bilateral septic knee joint prostheses from a distant cat bite. Ann Emerg Med. 1984;13(11):1065–1067. 83. Mellors JW, Schoen RT. Pasteurella multocida prosthetic joint infection. Ann Emerg Med. 1985;14(6):617. 84. Sugarman M, Quismorio FP, Patzakis MJ. Joint infection by Pasteurella multocida. Lancet. 1975;2(7947):1267. 85. Chikwe J, Bowditch M, Villar RN, et al. Sleeping with the enemy: Pasteurella multocida infection of a hip replacement. J R Soc Med. 2000;93(9):478–479.

44

Bone and Joint Infections

86. Guion TL, Sculco TP. Pasteurella multocida infection in total knee arthroplasty. Case report and literature review. J Arthroplasty. 1992;7(2):157–160. 87. Antuna SA, Mendez JG, Castellanos JL, et  al. Late infection after total knee arthroplasty caused by Pasteurella multocida. Acta Orthop Belg. 1997;63(4):310–312. 88. Braithwaite BD, Giddins G. Pasteurella multocida infection of a total hip arthroplasty. A case report. J Arthroplasty. 1992;7(3):309–310. 89. Arvan GD, Goldberg V. A case report of total knee arthroplasty infected by Pasteurella multocida. Clin Orthop. 1978(132):167–169. 90. Spagnuolo PJ. Pasteurella multocida infectious arthritis. Am J Med Sci. 1978;275(3): 359–363. 91. Maurer KH, Hasselbacher P, Schumacher HR. Letter: Joint infection by Pasteurella multocida. Lancet. 1975;2(7931):409. 92. Maradona JA, Asensi V, Carton JA, et al. Prosthetic joint infection by Pasteurella multocida. Eur J Clin Microbiol Infect Dis. 1997;16(8):623–625. 93. Griffin AJ, Barber HM. Joint infection by Pasteurella multocida. Lancet. 1975;1(7920): 1347–1348. 94. Chenouard R, Hoppe E, Lemarie C et al. A rare case of prosthetic joint infection associated with Coxiella burnetii. International J Infect Dis. 2019;87:166–169. 95. Tande AJ, Cunningham SA, Raoult D, et al. A case of Q fever prosthetic joint infection and description of an assay for detection of Coxiella burnetii. J Clin Microbiol 2014;51(1):66–69. 96. Million M, Bellevegue L, Labussiere AS ea. Culture‐negative prosthetic joint arthritis related to Coxiella burnetii. Am J Med. 2014;786. 97. Eldin C, Melenotte C, Mediannikov O. et al. From Q Fever to Coxiella Burnetii Infection: A Paradigm Change. Clin Microbiol Rev. 2017;30(1):115–190. 98. Meriglier E, Sungler A, Elsendoom A. et al. Osteoarticular manifestations of Q fever: A case series and literature review. Clin Microbiol Infect. 2018;24:8. 99. Weisenberg S, Perlada D, Peatman T. Q fever prosthetic joint infection. BMJ Case Rep. 2017:1–3. 100. Werno AM, Klena JD, Shaw GM, et al. Fatal case of Campylobacter lari prosthetic joint infection and bacteremia in an immunocompetent patient. J Clin Microbiol. 2002;40(3): 1053–1055. 101. Dumic I, Sengodan M, Franson J et al. Early onset prosthetic joint infection and bacteremia due to Campylobacter fetus subspecies fetus. Case Rep Infect Dis. 2017:1–6. 102. Peterson MC, Farr RW, Castiglia M. Prosthetic hip infection and bacteremia due to Campylobacter jejuni in a patient with AIDS. Clin Infect Dis. 1993;16(3):439–440. 103. Peterson MC, Farr RW, Castiglia M. Prosthetic hip infection and bacteremia due to Campylobacter jejuni in a patient with AIDS. Clin Infect Dis. 1993;16(3):439–440. 104. Zamora‐López MJ, Álvarez‐García P, García‐Campello M. Prosthetic hip joint infection caused by Campylobacter fetus: A case report and literature review. Rev Esp Quimioter. 2018;31(1):53–57. 105. Vasoo S, Scwab JJ, Cunningham SA. et al. Campylobacter prosthetic joint infection. J Clin Microbiol. 2014;52(5):1771–1774. 106. Lariviere LA, Gaudreau CL, Turgeon FF. Susceptibility of clinical isolates of Campylobacter jejuni to twenty‐five antimicrobial agents. J Antimicrob Chemother. 1986;18(6):681–685. 107. Endtz HP, Ruijs GJ, van Klingeren B, et al. Quinolone resistance in campylobacter isolated from man and poultry following the introduction of fluoroquinolones in veterinary medicine. J Antimicrob Chemother. 1991;27(2):199–208. 108. Tenover FC, Baker CM, Fennell CN et al. Antimicrobial resistance in Campylobacter species. In: Nachamkin I, Blaser MJ, Tomkins LS, editors. Campylobacter jejuni: current status and future trends. Washington, DC: American Society for Microbiology; 1992. p. 66–73. 109. Bates CJ, Clarke TC, Spencer RC. Prosthetic hip joint infection due to Campylobacter fetus. J Clin Microbiol. 1994;32(8):2037.

3  Unusual Microorganisms in Periprosthetic Joint Infection

45

110. Gazeau P, Rezig S, Quaesaet L, et al. Erysipelothrix rhusiopathiae knee prosthesis infection. Med Mal Infect. 2018;48(5):372–373. 111. Groeschel M, Forde T, Turvey S, et al. An unusual case of Erysipelothrix rhusiopathiae prosthetic joint infection from the Canadian Arctic: whole genome sequencing unable to identify a zoonotic source. BMC Infect Dis. 2019;19(1):282. 112. Hocqueloux L, Poisson DM, Sunder S, et al. Septic arthritis caused by Erysipelothrix rhusiopathiae in a prosthetic knee joint. J Clin Microbiol. 2010;48(1):333–335. 113. Traer EA, Williams MR, Keenan JN. Erysipelothrix rhusiopathiae infection of a total knee arthroplasty an occupational hazard. J Arthroplasty. 2008;23(4):609–611. 114. Troelsen A, Møller JK, Bolvig L, et al. Animal‐associated bacteria, Erysipelotrix rhusiopathiae, as the cause of infection in a total hip arthroplasty. J Arthroplasty. 2010;25(3):497.e21–23. 115. Brooke CJ, Riley TV. Erysipelothrix rhusiopathiae: bacteriology, epidemiology and clinical manifestations of an occupational pathogen. J Med Microbiol. 1999;48(9):789–799. 116. Al-Tawfiq JA, Babiker MM. Mulit-focal Clostridioides (Clostridium) difficile osteomyelitis in a patient with sickle cell anemia: case presentation and literature review. Diagn Microbiol Infect Dis. 220;96(1):114915. Doi: 10.1016/j.diagmicrobio.2019.114915. 117. Pron B, Merckx J, Touzet P, et al. Chronic septic arthritis and osteomyelitis in a prosthetic knee joint due to Clostridium difficile. Eur J Clin Microbiol Infect Dis. 1995;14(7):599–601. 118. Ranganath S, Midturi JK. Unusual case of prosthetic shoulder joint infection due to Clostridium difficile. Am J Med Sci. 2013;346(5):422–423. 119. McCarthy J, Stingemore N. Clostridium difficile infection of a prosthetic joint presenting 12 months after antibiotic‐associated diarrhoea. J Infect. 1999;39(1):94–96. 120. Brown ML, Drinkwater CJ. Hematogenous infection of total hip arthroplasty with Actinomyces following a noninvasive dental procedure. Orthopedics. 2012;35(7):e1086–1089. 121. Dagher R, Riaz T, Tande AJ, et al. Prosthetic Joint Infection due to Actinomyces species: A case series and review of literature. J Bone Jt Infect. 2019;4(4):174–180. 122. Dubourg G, Delord M, Gouriet F, et al. Actinomyces gerencseriae hip prosthesis infection: a case report. J Med Case Rep. 2015;9:223. 123. Hedke J, Skripitz R, Ellenrieder M, et al. Low‐grade infection after a total knee arthroplasty caused by Actinomyces naeslundii. J Med Microbiol. 2012;61(Pt 8):1162–1164. 124. Redmond SN, Helms R, Pensiero A. A Case of Actinomyces Prosthetic Hip Infection. Cureus. 2020;12(7):e9148. 125. Rieber H, Schwarz R, Krämer O, et  al. Actinomyces neuii subsp. neuii associated with periprosthetic infection in total hip arthroplasty as causative agent. J Clin Microbiol. 2009;47(12):4183–4184. 126. Sharma S, Sharma SC. Forgotten intrauterine contraceptive device ‐ A threat to total hip prosthesis: A case report with review of the literature. J Clin Orthop Trauma. 2016;7(2): 130–133. 127. Wu F, Marriage NA, Ismaeel A, et al. Infection of a total hip arthroplasty with actinomyces israelii: Report of a case. N Am J Med Sci. 2011;3(5):247–248. 128. Wust J, Steiger U, Vuong H, et al. Infection of a hip prosthesis by Actinomyces naeslundii. J Clin Microbiol. 2000;38(2):929–930. 129. Zaman R, Abbas M, Burd E. Late prosthetic hip joint infection with Actinomyces israelii in an intravenous drug user: case report and literature review. J Clin Microbiol. 2002; 40(11):4391–4392. 130. Valour F, Sénéchal A, Dupieux C, et al. Actinomycosis: etiology, clinical features, diagnosis, treatment, and management. Infect Drug Resist. 2014;7:183–197. 131. Steckelberg J, Osmon D. Prosthetic joint infections. In: Bisno A, Waldwogel FA, editors. Infections associated with indwelling medical devices. Washington, DC: ASM press; 2000. p. 259–290. 132. Stern SH, Sculco TP. Clostridium perfringens infection in a total knee arthroplasty. A case report. J Arthroplasty. 1988;3(Suppl):S37–40.

46

Bone and Joint Infections

133. Marchandin H, Jean‐Pierre H, Carriere C, et al. Prosthetic joint infection due to Veillonella dispar. Eur J Clin Microbiol Infect Dis. 2001;20(5):340–342. 134. Maniloff G, Greenwald R, Laskin R, et al. Delayed postbacteremic prosthetic joint infection. Clin Orthop. 1987(223):194–197. 135. Burnell CD, Turgeon TR, Hedden DR, et al. Paraneoplastic Clostridium septicum infection of a total knee arthroplasty. J Arthroplasty. 2011;26(4):666.e9–11. 136. Economedes DM, Santoro J, Deirmengian CA. Clostridium septicum growth from a total knee arthroplasty associated with intestinal malignancy: a case report. BMC Infect Dis. 2012;12:235. 137. Farer LS, Lowell AM, Meador MP. Extrapulmonary tuberculosis in the United States. Am J Epidemiol. 1979;109(2):205–217. 138. Watts HG, Lifeso RM. Tuberculosis of bones and joints. J Bone Joint Surg Am. 1996;78(2):288–298. 139. Neogi DS, Kumar A, Yadav CS, et  al. Delayed periprosthetic tuberculosis after total knee replacement: is conservative treatment possible? Acta Orthop Belg. 2009;75(1):136–140. 140. Berbari EF, Hanssen AD, Duffy MC, et al. Prosthetic joint infection due to Mycobacterium tuberculosis: a case series and review of the literature. Am J Orthop. 1998;27(3):219–227. 141. Spinner RJ, Sexton DJ, Goldner RD, et al. Periprosthetic infections due to Mycobacterium tuberculosis in patients with no prior history of tuberculosis. J Arthroplasty. 1996;11(2):217–222. 142. Amouyel T, Gaeremynck P, Gadisseux GB, et  al. Mycobacterium tuberculosis infection of reverse total shoulder arthroplasty: a case report. J Shoulder Elbow Surg. 2019;28:e271–e274. 143. Kim YH, Han DY, Park BM. Total hip arthroplasty for tuberculous coxarthrosis. J Bone Joint Surg Am. 1987;69(5):718–727. 144. Eskola A, Santavirta S, Konttinen YT, et al. Arthroplasty for old tuberculosis of the knee. Journal of Bone and Joint Surgery ‐ British Volume. 1988;70(5):767–769. 145. Santavirta S, Eskola A, Konttinen YT, et al. Total hip replacement in old tuberculosis A report of 14 cases. Acta Orthop Scand. 1988;59(4):391–395. 146. Bi AS, Li D, MA Y, et al. Mycobacterium tuberculosis as a cause of periprosthetic joint infection after total knee arthroplasty: a review of the literature. Cureus. 2019;11(3):e4325, 1–8. 147. Tokumoto JI, Follansbee SE, Jacobs RA. Prosthetic joint infection due to Mycobacterium tuberculosis: report of three cases. Clin Infect Dis. 1995;21(1):134–136. 148. American Thoracic Society, CDC, IDSA. Treatment of tuberculosis. MMWR. 2003;52(RR 11):57. 149. Uhel F, Corvaisier G, Poinsignon Y, et al. Mycobacterium tuberculosis prosthetic joint infections: a case series and literature review. J Infect 2019;78:27–34. 150. Wolfgang GL. Tuberculosis joint infection following total knee arthroplasty. Clin Orthop. 1985(201):162–166. 151. Ueng WN, Shih CH, Hseuh S. Pulmonary tuberculosis as a source of infection after total hip arthroplasty A report of two cases. Int Orthop. 1995;19(1):55–59. 152. Krappel FA, Harland U. Failure of osteosynthesis and prosthetic joint infection due to Mycobacterium tuberculosis following a subtrochanteric fracture: a case report and review of the literature. Arch Orthop Trauma Surg. 2000;120(7‐8):470–472. 153. Kreder HJ, Davey JR. Total hip arthroplasty complicated by tuberculous infection. J Arthroplasty. 1996;11(1):111–114. 154. McCullough CJ. Tuberculosis as a late complication of total hip replacement. Acta Orthop Scand. 1977;48(5):508–510. 155. Johnson R, Barnes KL, Owen R. Reactivation of tuberculosis after total hip replacement. Journal of Bone and Joint Surgery ‐ British Volume. 1979;61‐B(2):148–150. 156. Leach WJ, Halpin DS. Mycobacterium bovis infection of a total hip arthroplasty: a case report. J Bone Joint Surg Br. 1993;75(4):661–662. 157. Gomez E, Chiang T, Louie T, et al. Prosthetic Joint Infection due to Mycobacterium bovis after Intravesical Instillation of Bacillus Calmette‐Guerin (BCG). Int J Microbiol. 2009;2009:527208.

3  Unusual Microorganisms in Periprosthetic Joint Infection

47

158. Langlois ME, Ader F, Dumistrescu O, et al. Mycobacterium bovis prosthetic joint infection. Med Mal Infect. 2016;46(8):445–448. 159. Metayer B, Menu P, Khatchatourian L, et al. Prosthetic joint infection with pseudo‐tumoral aspect due to Mycobacterium bovis infection after Bacillus‐Calmette‐Guerin therapy. Ann Phys Rehabil Med. 2018;61(1):62–64. 160. Nguyen MH, Giordani MM, Thompson GR, 3rd. The double‐edged sword ‐ prosthetic joint infection following BCG treatment for bladder cancer: a case report. BMC Infect Dis. 2019;19(1):331. 161. Patel A, Elzweig J. Mycobacterium bovis prosthetic joint infection following intravesical instillation of BCG for bladder cancer. BMJ Case Rep. 2019;12(12). 162. Storandt M, Nagpal A. Prosthetic joint infection: an extremely rare complication of intravesicular BCG therapy. BMJ Case Rep. 2019;12(12). 163. Williams A, Arnold B, Gwynne‐Jones DP. Mycobacterium bovis infection of total hip arthroplasty after intravesicular Bacillus Calmette‐Guérin. Arthroplast Today. 2019;5(4):416–420. 164. Chazerain P, Desplaces N, Mamoudy P, et al. Prosthetic total knee infection with a bacillus Calmette Guerin (BCG) strain after BCG therapy for bladder cancer. J Rheumatol. 1993;20(12):2171–2172. 165. Badelon O, David H, Meyer L, et al. Mycobacterium fortuitum infection after total hip prosthesis. A report of 3 cases. Rev Chir Orthop Reparatrice Appar Mot. 1979;65(1):39–43. 166. Booth JE, Jacobson JA, Kurrus TA, et  al. Infection of prosthetic arthroplasty by Mycobacterium fortuitum. Two case reports. J Bone Joint Surg Am. 1979;61(2):300–302. 167. Herold RC, Lotke PA, MacGregor RR. Prosthetic joint infections secondary to rapidly growing Mycobacterium fortuitum. Clin Orthop. 1987(216):183–186. 168. Horadam VW, Smilack JD, Smith EC. Mycobacterium fortuitum infection after total hip replacement. South Med J. 1982;75(2):244–246. 169. Moerman J, Vandepitte J, Corbeel L, et al. Iatrogenic infections caused by the Mycobacterium fortuitum‐chelonei complex. Report of two cases and review. Acta Clin Belg. 1985;40(2): 92–98. 170. Pring M, Eckhoff DG. Mycobacterium chelonae infection following a total knee arthroplasty. J Arthroplasty. 1996;11(1):115–116. 171. Heathcock R, Dave J, Yates MD. Mycobacterium chelonae hip infection. J Infect. 1994;28(1):104–105. 172. Amit P, Rastogi S, Marya S. Prosthetic knee joint infection due to Mycobacterium abscessus. Indian J Orthop. 2017;51(3):337–342. 173. Eid AJ, Berbari EF, Sia IG, et al. Prosthetic joint infection due to rapidly growing mycobacteria: report of 8 cases and review of the literature. Clin Infect Dis. 2007;45(6):687–694. 174. Napaumpaiporn C, Katchamart W. Clinical manifestations and outcomes of musculoskeletal nontuberculous mycobacterial infections. Rheumatol Int. 2019;39(10):1783–1787. 175. Nengue L, Diaz MAA, Sherman CE, et  al. Mycobacterium abscessus Prosthetic Joint Infections of the Knee. J Bone Jt Infect. 2019;4(5):223–226. 176. Pace V, Antinolfi P, Borroni E, et al. Treating Primary Arthroprosthesis Infection Caused by Mycobacterium abscessus subsp. abscessus. Case Rep Infect Dis. 2019;2019:5892913. 177. Petrosoniak A, Kim P, Desjardins M, et al. Successful treatment of a prosthetic joint infection due to Mycobacterium abscessus. Can J Infect Dis Med Microbiol. 2009;20(3):e94–96. 178. Wang SX, Yang CJ, Chen YC, et al. Septic arthritis caused by Mycobacterium fortuitum and Mycobacterium abscessus in a prosthetic knee joint: case report and review of literature. Intern Med. 2011;50(19):2227–2232. 179. Spanyer J, Foster S, Thum‐DiCesare JA et  al Mycobacterium abscessus: A Rare Cause of Periprosthetic Knee Joint Infection. MD Edge Surgery. 2018. 180. Wallace RJ, Jr, Swenson JM, Silcox VA, et al. Disk diffusion testing with polymyxin and amikacin for differentiation of Mycobacterium fortuitum and Mycobacterium chelonei. J Clin Microbiol. 1982;16(6):1003–1006.

48

Bone and Joint Infections

181. Swenson JM, Wallace RJJ, Silcox VA, et al. Antimicrobial susceptibility of five subgroups of Mycobacterium fortuitum and Mycobacterium chelonae. Antimicrob Agents Chemother. 1985;28:807. 182. Martin ML, Dall L. Emergence of multidrug‐resistant Mycobacterium fortuitum during treatment. Chest. 1984;85(3):440–441. 183. Malhotra R, Kiran B, al. Gautam D, et  al. Mycobacterium abscessus periprosthetic joint infection following bilateral total knee arthroplasty. ID Cases. 2019;17:e00542. 184. Wallace RJ, Jr, Wiss K, Bushby MB, et al. In vitro activity of trimethoprim and sulfamethoxazole against the nontuberculous mycobacteria. Rev Infect Dis. 1982;4(2):326–331. 185. Wallace RJ, Jr, Jones DB, Wiss K. Sulfonamide activity against Mycobacterium fortuitum and Mycobacterium chelonei. Rev Infect Dis. 1981;3(5):898–904. 186. Silcox VA, Good RC, Floyd MM. Identification of clinically significant Mycobacterium fortuitum complex isolates. J Clin Microbiol. 1981;14(6):686–691. 187. Brown‐Elliott BA, Wallace RJ, Jr. Clinical and taxonomic status of pathogenic nonpigmented or late‐pigmenting rapidly growing mycobacteria. Clin Microbiol Rev. 2002;15(4):716–746. 188. Wallace RJ, Jr, Brown‐Elliott BA, Ward SC, et al. Activities of linezolid against rapidly growing mycobacteria. Antimicrob Agents Chemother. 2001;45(3):764–767. 189. Howard ST, Rhoades E, Recht J, et al. Spontaneous reversion of Mycobacterium abscessus from a smooth to a rough morphotype is associated with reduced expression of glycopeptidolipid and reacquisition of an invasive phenotype. Microbiology. 2006;152(Pt 6):1581–1590. 190. Gupta A, Clauss H. Prosthetic joint infection with Mycobacterium avium complex in a solid organ transplant recipient. Transpl Infect Dis. 2009;11(6):537–540. 191. Ingraham NE, Schneider B, Alpern JD. Prosthetic Joint Infection due to Mycobacterium avium‐intracellulare in a Patient with Rheumatoid Arthritis: A Case Report and Review of the Literature. Case Rep Infect Dis. 2017;2017:8682354. 192. McLaughlin JR, Tierney M, Harris WH. Mycobacterium avium intracellulare infection of hip arthroplasties in an AIDS patient. J Bone Joint Surg Br. 1994;76(3):498–499. 193. Tan EM, Marcelin JR, Mason E, et al. Mycobacterium avium intracellulare complex causing olecranon bursitis and prosthetic joint infection in an immunocompromised host. J Clin Tuberc Other Mycobact Dis. 2016;2:1–4. 194. Isono SS, Woolson ST, Schurman DJ. Total joint arthroplasty for steroid‐induced osteonecrosis in cardiac transplant patients. Clin Orthop. 1987(217):201–208. 195. Spinner RJ, Sexton DJ, Vail TP. Mycobacterium avium intracellulare infection. J Bone Joint Surg Br. 1995;77(1):165. 196. Horsburgh CR, Jr. Mycobacterium avium complex infection in the acquired immunodeficiency syndrome. N Engl J Med. 1991;324(19):1332–1338. 197. Griffith DE, Aksamit T, Brown‐Elliott BA, et al. An official ATS/IDSA statement: diagnosis, treatment, and prevention of nontuberculous mycobacterial diseases. Am J Respir Crit Care Med. 2007;175(4):367–416. 198. Han Z, Burnham C‐A, Clohisy J, et al. Mycoplasma pneumoniae Periprosthetic Joint Infection Identified by 16S Ribosomal RNA Gene Amplification and Sequencing: A Case Report. The Journal of bone and joint surgery American volume. 2011;93:e103. 199. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et al. A Novel Prosthetic Joint Infection Pathogen, Mycoplasma salivarium, Identified by Metagenomic Shotgun Sequencing. Clin Infect Dis. 2017;65(2):332–335. 200. Sneller M, Wellborne F, Barile MF, et al. Prosthetic joint infection with Mycoplasma hominis. J Infect Dis. 1986;153(1):174–175. 201. Madoff S, Hooper DC. Nongenitourinary infections caused by Mycoplasma hominis in adults. Rev Infect Dis. 1988;10(3):602–613. 202. Rieber H, Frontzek A, Fischer M. Periprosthetic joint infection associated with Mycoplasma hominis after transurethral instrumentation in an immunocompetent patient. Unusual or underestimated? A case report and review of the literature. Int J Infect Dis. 2019;82:86–88.

3  Unusual Microorganisms in Periprosthetic Joint Infection

49

203. Kenny GE, Cartwright FD. Susceptibilities of Mycoplasma hominis, M. pneumoniae, and Ureaplasma urealyticum to GAR‐936, dalfopristin, dirithromycin, evernimicin, gatifloxacin, linezolid, moxifloxacin, quinupristin‐dalfopristin, and telithromycin compared to their susceptibilities to reference macrolides, tetracyclines, and quinolones. Antimicrob Agents Chemother. 2001;45(9):2604–2608. 204. Roerdink RL, Douw CM, Leenders AC, et  al. Bilateral periprosthetic joint infection with Ureaplasma urealyticum in an immunocompromised patient. Infection. 2016;44(6):807–810. 205. Rouard C, Pereyre S, Abgrall S, et al. Early prosthetic joint infection due to Ureaplasma urealyticum: Benefit of 16S rRNA gene sequence analysis for diagnosis. J Microbiol Immunol Infect. 2019;52(1):167–169. 206. Sköldenberg OG, Rysinska AD, Neander G, et al. Ureaplasma urealyticum infection in total hip arthroplasty leading to revision. J Arthroplasty. 2010;25(7):1170.e11–13. 207. Voutsinas S, Sayakos J, Smyrnis P. Echinococcus infestation complicating total hip replacement. A case report. J Bone Joint Surg Am. 1987;69(9):1456–1458. 208. Notarnicola A, Panella A, Moretti L, et al. Hip joint hydatidosis after prosthesis replacement. Int J Infect Dis. 2010;14 Suppl 3:e287–290. 209. Perlick L, Sommer T, Zhou H, et al. Atypical prosthetic loosening in the hip joint. Radiologe. 2000;40(6):577–579. 210. Aktan AO, Yalin R. Preoperative albendazole treatment for liver hydatid disease decreases the viability of the cyst. Europ J Gastroenterol Hepatol. 1996;8(9):877–879. 211. Fresard A, Guglielminotti C, Berthelot P, et al. Prosthetic joint infection caused by Tropheryma whippelii (Whipple’s bacillus). Clin Infect Dis. 1996;22(3):575–576. 212. Cremniter J, Bauer T, Lortat‐Jacob A, et al. Prosthetic hip infection caused by Tropheryma whipplei. J Clin Microbiol. 2008;46(4):1556–1557. 213. Adrados M, Wiznia DH, Golden M, et al. Lyme periprosthetic joint infection in total knee arthroplasty. Arthroplasty Today. 2018;4(2):158–161. 214. Collins KA, Gotoff JR, Ghanem ES. Lyme Disease: A Potential Source for Culture‐negative Prosthetic Joint Infection. J Am Acad Orthop Surg Glob Res Rev. 2017;1(5):e023. 215. Wright WF, Oliverio JA. First Case of Lyme Arthritis Involving a Prosthetic Knee Joint. Open Forum Infect Dis. 2016;3(2):ofw096.

Chapter 4

Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques Maria Eugenia Portillo and Stéphane Corvec

Introduction Despite the correct implementation of special diagnostic culture techniques, such as tissue sample processing with beadmill, prolonged incubation time, or sonication of removed implants, a considerable number of bone and joint infections (BJI) are either culture‐negative or misjudged as aseptic failure [1–3]. Misinterpretation may lead to wrong or needless antimicrobial treatment, or even to unnecessary surgery [4,5]. The diagnostic value of non‐culture techniques has been investigated in synovial fluid, periprosthetic tissue samples (homogenized manually or by semi‐automated techniques [6]), and sonication fluid [7–9]. Although these innovative techniques have proved to be helpful, strict conditions have to be respected by specifically trained staff in order to avoid any contamination. Different practicable improvements, such as real‐time technology, automated systems, and commercial kits are now available, and support the microbiologists in daily practice [10]. In this chapter, we describe non‐culture techniques based on nucleic acid amplification, sequencing, and mass‐spectrometry methods. Polymerase chain reaction (PCR) is a frequently used technique in most microbiology laboratories that allows detection of a nucleic acid fragment by amplifying a sequence. There are different PCR types that can be used in the BJI diagnosis, and each one requires the selection of appropriate primers (Table 4.1).

Broad‐Range PCR The value of broad‐range PCR has been extensively studied in the diagnosis of BJI, but extraction of pathogen DNA from bone, joint tissue, or implant remains challenging [11].

Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

51

52

Bone and Joint Infections

Table 4.1.  Different types of PCR applied in the diagnosis of BJI. Simultaneous amplification and detection

PCR

General description

Quantification

Broad‐ range PCR

Detection of genes universally present in microorganisms which requires posterior sequencing step for identification

No

Semiquantitative by comparing gel band intensity

Targeted PCR

Specific detection of a particular microorganism and/or resistance mechanism

No: Gel‐based PCR Yes: Real‐time PCR

Semiquantitative Yes

Multiplex PCR

Simultaneously detection of several microorganisms and/or resistance mechanisms by adding the sets of primers of interest

No: Gel‐based PCR Yes: Real‐time PCR

Semiquantitative Yes

Disruption of the biofilm is an essential step to release the DNA in order to improve the sensitivity of broad‐range bacterial PCR [12]. Pathogens can be identified even if genus or species are unknown using universal primers to amplify bacterial or fungal DNA, followed by the identification of the species by sequencing, a technique which is also called universal PCR. This methodology is applied in isolated strains that are difficult to identify using conventional techniques. More recently, it has also been used directly from clinical samples, where the detection and identification of the pathogen by conventional techniques remains difficult or not possible [11]. Broad‐range PCR consists of two steps: the amplification of the bacterial or fungal DNA within the sample, and the subsequent sequencing of the PCR fragment for the identification of the microorganism (Figure 4.1). The regions of the genome that are used must fulfill fundamental characteristics. First, they must be present in all bacterial or fungal species; second, they should contain highly conserved sequences to which the primers are directed; and finally, they have to include polymorphic sequences, in order to distinguish different species. After amplifying and sequencing the fragment, the obtained sequence is compared to those placed in public databases such as NCBI GenBank. For sequence alignment, programs such as BLAST are available, and allow online sequence comparison [13]. In bacteria, species identification at the molecular level is based on analysis of the 16S rRNA gene sequence. This molecule is 1,500 bp in size. Usually, it is sufficient to analyze the first 500 bp of this gene, since it is the most variable region, but in other cases, the entire gene must be sequenced, or even the study of other sites in the genome must be used. Similarly, for fungal broad‐range PCR, primers targeting conserved areas of genes encoding the 18S, 5.8S, and 28S ribosomal subunits are used. ITS‐1 and 2 regions are among these genes, which allow discriminating between the different fungal species [14]. Therefore, broad‐range PCR allows the identification of microorganisms previously not thought to cause infection, despite it being less sensitive than targeted or multiplex PCR. The main disadvantages of broad‐range PCR are lack of sensitivity, false‐positive results resulting from contamination, need of subsequent sequencing, and challenge of result interpretation [15]. Also, the sequence analysis may be uninformative and thus, this approach may not indicate whether a polymicrobial or monomicrobial infection is present (i.e., due to overlapping electropherogram peaks) or misleading (i.e., due to missed detection of minority species). This issue may be mitigated by more detailed sequence

4  Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques

1

53

2

Sample Marker

1500 bp

Figure 4.1.  Broad‐range PCR scheme for bacteria and fungi. From the corresponding genome region: 1) a DNA fragment is amplified using primers targeting conserved regions present in all (universal) bacteria or fungi and 2) polymorphic regions are sequenced 3) to identify the species. Source: ABI chromatogram and National Institutes of Health.

analysis, or through software manipulation of sequencer data (i.e. use of the software RipSeq) [16]. During a prospective multicenter cross sectional study, Bémer et  al. [17] demonstrated a sensitivity of only 73.3% in the diagnosis of prosthetic joint infections (PJI). Although new PCR assays have been developed, the sensitivity of PCR for the diagnosis of PJI is variable in different studies, whereas its specificity is reliably high, allowing the exclusion of PJI [18].

Targeted PCR Targeted or specific PCR can be developed for any known microorganism, and can be designed to be extremely sensitive. The analysis is typically performed in real‐time, because the amplification process and detection occur simultaneously in the same closed vial. Furthermore, it is possible to measure the amount of DNA synthesized at each moment during amplification, since the fluorescence emission produced in the reaction is proportional to the amount of formed DNA. Therefore, real‐time PCR has three important advantages compared to conventional or agarose gel‐based PCR, namely higher speed, lower contamination rate, and more precise DNA quantification. Likewise, point mutations that may be related to antimicrobial resistance or virulence factors can also be detected. However, if the gene needs to be sequenced, it is better to perform an agarose gel‐based PCR [19,20].

54

Bone and Joint Infections

Different genes have been used as target to accurately identify most coagulase‐negative staphylococci (CNS) implicated in BJI based on PCR amplification and sequencing. A genetic classification and identification of Staphylococcus species was conducted based on the comparison of partial glyceraldehyde‐3‐phosphate dehydrogenase gene, 16S rRNA, hsp60, rpoB, sodA, and tuf gene sequences [21]. Kingella kingae is currently recognized as the prime etiology of BJI in children aged 6–48 months (see Chapter 8). The organism is fastidious, its growth is inhibited by synovial fluid and bone exudates, and its presence in clinical specimens is commonly missed by traditional culture methods. Therefore, the use of specific PCR is highly recommended in septic arthritis and osteomyelitis in infants and young children [22] (Figure 4.2). Different companies have developed an automated, easy‐to‐use, fast, and accurate real‐time Staphylococcus aureus identification PCR which can be combined with the search for methicillin resistance (mecA and/or mecC gene) to optimize the management and the appropriate treatment of the patient, especially in acute septic arthritis cases [23]. The addition of specific nucleic acid amplification testing is also suggested if bacterial cultures do not reveal growth or if fastidious microorganisms are suspected. This may typically occur in the case of vertebral osteomyelitis (Brucella spp., Mycobacterium tuberculosis, Tropheryma whipplei, Coxiella spp., or Mycoplasma spp.) [24] (see Chapters 18 and 19).

Multiplex PCR Multiplex PCR is a technique in which more than two sets of primers are involved in the process of amplifying various target sequences, allowing the simultaneous detection and identification of different genes. The main advantage of these systems is the ability of

A

B

Molecular marker

Sample

175 bp

Figure 4.2.  Targeted PCR of cpn60 gene from Kingella kingae detected A) in conventional electrophoresis gel and B) in real‐time PCR.

4  Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques

55

grouping in a single process different targeted PCR, simplifying the process, saving time and cost, as well as shortening the diagnostic time [5]. In a recent meta‐analysis, it has been shown that multiplex PCR from sonication fluid of prosthetic implants is reliable and of great value for the diagnosis of BJI [25]. Several groups have investigated different multiplex PCR panels. However, these tests have mainly been developed for bloodstream infections. Thus, their use for the rapid diagnosis of BJI is off‐label [26–28]. There are different primer panels (Table 4.2), including primers for low‐virulent organisms frequently involved in chronic or delayed PJI, such as Corynebacterium spp., Cutibacterium spp., or other anaerobes.

Next‐Generation Sequencing Approach Beside culturomics methods, in the last two decades different non‐culture techniques have been developed to better detect microorganisms in clinical perioperative samples, either bacteria or fungi. In order to improve the BJI diagnosis, different teams have developed original sequencing approaches. Henceforth, with the reduced cost of sequencing and the availability of next‐generation sequencing (NGS) methods, it is reasonable to capture all the potential sequences which represent the whole diversity of pathogens present in a joint, a bone biopsy, or a tissue in contact with the device (Table 4.3). Interestingly, compared with the targeted 16S universal PCR called the metataxonomic method, the shotgun random approach allowed the catching of all the sequences including those from fungi or potentially viruses (Table 4.4). In 2018, Dekker [29] wrote that metagenomics approaches are close to reality for the diagnostic workup of clinical infectious diseases. However, we need to highlight that the first challenge is a potential microbial DNA contamination, which may occur either during sampling, linked to the reagents, due to contaminated instrument surfaces, or in the environment [30]. Therefore, the interpretation of results needs strict criteria and clinical knowledge in order to detect fastidious microorganisms and to rule out contaminants. The specialized microbiologists should assist clinicians in interpreting the results of NGS in interdisciplinary meetings. For meaningful use, this new tool should only be performed in patients with chronic infection, with culture‐negative Table 4.2.  Multiplex PCR platforms used in BJI. Platform

Manufacturer

BJI cartridge

Target

Resistance genes

Unyvero i60

Curetis

Yes

GPBa including Corynebacterium spp., Granulicatella spp., and Abiotrophia spp., GNBb including non‐fermenting, anaerobes, and fungi

Yes

SeptiFastc

Roche

No

GPBa, GNBb including non‐fermenting and fungi

Yes

FilmArray

BioFire (BJI)

Yes

GPBa, GNBb including non‐fermenting, Kingella kingae, anaerobes and fungi

Yes

GPB Gram‐positive bacteria GNB Gram‐negative bacteria c Soon discontinued a

b

56

Bone and Joint Infections

Table 4.3.  Different applications of Next‐Generation Sequencing (NGS) in clinical microbiology. Methodology

Applications

Whole Genome Sequencing

WGS of a pure organism isolated from culture.

Targeted NGS directly from specimen

16S rDNA from a clinical specimen for bacterial profiling or PCR amplification of other specified targets followed by sequencing.

Metagenomic NGS directly from specimen

Nucleic acid composition of the specimens includes host, microbiome, and the maybe the accidental introduction of contaminating nucleic acid.

Table 4.4.  Terms used in the area of Next‐Generation Sequencing. Technology

Definitions

Next‐Generation Sequencing

High‐throughput, massively parallel sequencing of DNA fragments independently and simultaneously.

Whole Genome Sequencing (WGS)

Sequencing of a microbial genome. WGS can be applied to pure culture growth or directly from specimen.

Targeted WGS

Uses a selection process to enrich for specific targets before sequencing.

Metagenomic WGS

Sequencing of all nucleic acids detected directly from patient specimens.

specimens, or with current or recent antimicrobial therapy. Another issue is the frequent overrepresentation of human host DNA (neutrophils, osteoblast and osteoclast cells) which interferes with the microorganisms’ sequences, less frequent in terms of amount or proportion. For these reasons, we should still be reserved with this novel technique. However, NGS‐based technology will gradually improve the etiologic diagnosis of BJI. They will reach in the near futures the standard of care in infectious disease diagnostics. In 2017, Street et al. [31] from the Oxford group published one of the first studies about NGS (metagenomic sequencing) on 97 sonication fluid samples from various orthopedic devices. They concluded that this technology is a promising method, especially if in the future there will be improved techniques to avoid microbial, as well as human DNA, contamination. The increasing availability of different systems, which are quicker than culture and even portable, will offer a new clinically useful rapid diagnostic tool (Table 4.3). However, they also highlighted the risk of false‐positive results, especially with Cutibacterium acnes, a bacterium from the skin microbiome, or with environmental bacteria linked to water. In addition, false‐negative results may occur, due to a limited database. Therefore, each case should be discussed in a multidisciplinary team [31]. During the same year, Thoendel et al. [32] from Robin Patel’s group reported the usefulness of this approach to identify a wide range of BJI pathogens, including difficult‐to‐detect pathogens, especially in culture‐negative infections. Among others, Mycoplasma species, anaerobes including Cutibacterium acnes, fungi, Staphylococcus spp., and Streptococcus spp. were identified as new microorganisms in culture‐negative BJI. Mixed and polymicrobial PJI could also be detected. However, the data interpretation needs attention. Read counts and depth of genome coverage are important to distinguish non‐cultured potential microorganisms detected from uninfected cases or negative controls. This constitutes a major challenge and the background reads should be analyzed carefully to determine their relevance.

4  Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques

57

During the last couple of years, different case reports or outbreaks have been reported in the literature using this new molecular strategy in the orthopedic field [33–37]. Furthermore, other samples such as synovial fluids have been tested using NGS. The metagenomic shotgun sequencing approach can detect microorganisms causing BJI, and might be useful especially for negative‐culture cases when applied to synovial fluids. Ivy et al. [38] described the workflow methodology and interpretation performed during their study. They present a list of the most frequent contaminant microorganisms depending on the read counts. The read counts went from 2,669 in Cutibacterium to 2,017,563 in Acinetobacter. Despite the lack of a generally recognized gold standard in all studies analyzing with NGS technology in BJI, it seems important to define a threshold depending on the reads detected. Moreover, to limit this bias between pathogens and contaminating microbial DNA, an effective microbial enrichment and DNA isolation remain crucial to allow better analysis of the samples. Thus, using a strict protocol and prudent interpretation allows us to correctly recognize contaminants. In this field, another technology approach is available. Sanderson et al. [39] demonstrated the proof of concept using the long‐read sequencing MinION technology on 9 specimens (7 positive and 2 negative). Despite almost 90% of human DNA, they managed to detect bacterial infections from DNA extracted directly from sonication fluid samples, and potentially provide an answer within minutes of starting sequencing. One case was polymicrobial with three different microorganisms, of which Fusobacterium nucleatum was not detected by tissue biopsy or sonication fluid culture [39]. Thus, this strategy, which is not available in all laboratories, may provide improvement in the diagnostic process in special cases. However, a definition of patient eligibility criteria is needed. If used in the correct population, the patient outcome may improve, due to the shorter diagnostic delay, which results in a more rapid targeted antibiotic treatment. Regarding the analysis of polymicrobial infections using a 16S metagenomic approach, Chen et  al. [40] demonstrated that the three databases used provided similar results. Indeed, the only variability observed was noticed on less abundant genera, especially those for which the relative abundance was less than 5% of the total reads. One limit of this recent study is the low number of patients (n=11) and maybe the primers selected, as previously mentioned [40]. At last, with the democratization of NGS (metagenomic and/or shotgun approaches), and the decreased costs, it would be interesting to standardize the method used to improve the BJI diagnosis, especially using NGS (Figure 4.3). Indeed, as the experimental design and analysis methods of each study are different, it would be logical to establish an NGS gold standard for PJI constituing a suitable standard procedure for perioperative specimens, which allows comparing of different studies. Thus, recently Li et al. [41] published a review and meta‐analysis of the performance of sequencing assays in the diagnosis of PJI. Even if it sometimes remains difficult to compare different studies, sequencing assays have the potential to improve biological diagnosis of PJI, especially when culture results are negative. However, one should be cautious about which type of culture is performed. According to our experience, optimized culture after beadmill processing remains better than other methods. The authors highlighted the importance of an antibiotic‐free interval to improve the performances and enhance the ability to detect microorganisms, especially those difficult to culture but susceptible to antibiotics. They also highlighted the superiority of sequencing by synthesis. Although Sanger sequencing remains more easily available and accurate in sequencing, NGS might be superior in consideration of cost and promptness [41]. At last in their model, Torchia et al. [42] suggested that NGS should be reserved for patients with a high pretest probability of PJI and specific clinical c­ ontexts,

58

Bone and Joint Infections

Figure 4.3.  Next‐Generation Sequence approaches, including specimen preparation and bioinformatics analysis.

such as antibiotic pretreatment, repetitive previous interventions, risk of polymicrobial infection, and chronic infection. We clearly need to define the indications and subgroups of patients for which NGS offers clinical benefit [42]. Recently, Wang et al. [43] concluded that targeted antibiotic treatment based on NGS results is reliable to identify pathogens in patients with culture‐negative PJI, and that it yields a favorable outcome in a short period of time. Therefore, we can ask two questions. First, should this technology be included in the diagnostic routine procedure? Second, does NGS indeed improve the outcome of patients with culture‐negative PJI due to the earlier switch from empiric to adapted antibiotic treatment? This may be especially important in patients with fastidious microorganisms or those with polymicrobial PJI, because adapted therapy may be delayed up to one week in this population, even with optimal culture methods [44]. Taken together, the NGS approach will be an additional diagnostic test when standard of care testing is uninformative. However, in the near future, it will still be limited to specialized laboratories, where a sequencing platform (Table 4.5) and a bioinformatic pipeline are available, and the lab workers have the required ability. It is probable that in the near future, the costs for this technology will decrease, allowing a broader use of the NGS approach.

Mass Spectrometry‐Based Methods The increase of this new technology to identify microorganisms (MALDI‐TOF spectrometry: Matrix Assisted Laser Desorption Ionization – Time of Flight) more than a decade ago in microbiology has constituted a real evolution and improvement in the turnaround time of identification [45]. With this method, the identification process took 24h less time, when the culture was positive. Quickly, its use has been diverted to obtain a rapid identification directly from samples, in particular blood culture bottles but also

4  Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques

59

Table 4.5.  Current common NGS platforms used for the diagnosis of clinical microbiology. Platforms

Manufacturer

NSG Generation

Technology

iSeq,MiSeq…

Illumina

Second

Sequencing by synthesis

Ion Torrent

ThermoFisher

Second

Sequencing by synthesis

MinION, GridION…

Oxford Nanopore

Third

Exonuclease sequencing and detection by electrical conductivity

Sequel, RSII

Pacific Biosiences

Third

Sequencing by synthesis

Figure 4.4.  Performance of MALDI‐TOF analysis.

later in the synovial fluids [46,47]. For more than a decade, the majority of the microbiology labs have been using the MALDI‐TOF technology. Directly from the agar plate, this identification method allows a rapid diagnostic (Figure 4.4) [5]. In 2010, Harris et al. [48] reported the ability and the great potential of the MALDI Biotyper software to recognize clonally related strains within a species group (i.e. sub‐­ typing). This MALDI‐TOF/MS MALDI Biotyper system provides a promising rapid and reliable method of identifying clinical isolates from BJI to the species level, and has potential for sub‐typing [48]. Peel et al. [49] confirmed these data for the diagnosis of PJI. The likelihood that a microorganism was a pathogen or contaminant differed with the prosthetic joint location, particularly in the case of Cutibacterium acnes, which is especially frequent in shoulder PJI. This spectral method constitutes a valuable tool for the identification of bacteria isolated from prosthetic joints, providing species‐level identification that may inform culture interpretation of pathogens versus contaminants [49]. In the near future, it can be also used with adapted software to screen more colonies and to distinguish monoclonal or polyclonal infections due to the same species but potential different clusters [50]. Lallemand et al. [51] assessed the opportunity to perform MALDI‐TOF identification directly on bone samples for comparison of the different methods available or used

60

Bone and Joint Infections

in the routine lab. The use of MALDI‐TOF directly on the specimen provided no useful contribution in the routine diagnosis, potentially due to a low bacterial burden in operating room samples for BJI diagnosis. The highest sensitivity was obtained with optimized culture (85.9%). Direct examination remains insensitive (31.7%) while ­ MALDI‐TOF identification represented a 6.3% sensitivity [51]. This method can be improved after extraction following enrichment in blood culture bottles, as previously evoked. Use of MALDI‐TOF identification on enrichment pellets of bone samples is an accurate, rapid, and robust method for bacterial identification of clinical isolates from BJI, except for streptococci, whose identification to species level remains difficult as it has been previously reported [47]. On the other hand, Greenwood‐Quaintance et  al. [52] reported a study on the assessment of a new technology called PCR‐electrospray ionization mass spectrometry (PCR‐ESI/MS) after the first study by Jacovides et  al. [53] based on a research use‐only platform with an old version of the software. In fact, it combined a broad‐ range PCR with a MALDI‐TOF technology. Obviously, the database is different and should be regularly updated. They compared PCR‐ESI/MS results to culture using sonicate fluids from 431 explanted‐knee or ‐hip patients. Briefly, one ml of non‐ concentrated sonicate fluid was thawed and DNA extraction was performed before using the PLEX‐ID PCR‐ESI/MS instrument manufactured by Abbott. The sensitivity for detecting BJI were 77.6% for PCR‐ESI/MS and 69.7% for culture (p=0.0105) and the specificities were 93.5 and 99.3%, respectively (p=0.0002). More sensitive but less specific, this method seems to be a useful tool for rapid detection of BJI and can integrate the arsenal of methods used to reach the causing bacteria. The same group from the Mayo clinic published a similar study, using synovial fluid specimens [54]. They reported an 81% sensitivity and a 95% specificity for the diagnosis of PJIs. Like with other methods, it is necessary to have an appropriate cut‐off for meaningful interpretation. Other groups have reported the utility of this methodology for detecting DNA from hundreds of different microorganisms in whole blood [55]. The detection of the microorganism is based on the system’s software and according to two parameters, a Q‐score ranging from 0 (low) to 1 (high) representing a relative measure of the strength (up to 0.9 is reported) and the level of detection representing a semiquantitative measure of the amount of amplified DNA calculated relative to an internal caliper corresponding to genome equivalents. These parameters should be well controlled in the interpretation of the data, in order to avoid misinterpretation of false‐positive or false‐negative results. Thus, this method can detect with a level of 15 and a Q‐score of 0.9 a Micrococcus luteus, but it does not mean infection. Indeed, in Jacovides study [53], a high rate of false positivity was highlighted with 49 out of 57 aseptic failure patients having positive PCR‐ESI/MS results. Again, like with the NGS, the microorganism detected, the values found for the interpretation, the clinic, and the multidisciplinary approach remain the keys to success for the diagnosis and the treatment of a BJI. Moreover, like with conventional molecular techniques or NGS, microbial DNA detection should not be interpreted systematically as infection. DNA persistence in clinical samples, bone biopsy, synovial fluid, or tissue following successful treatment is real. Bémer et al. [56] reported the detection by 16S PCR of Listeria monocytogenes in a perfectly healthy patient, in whom antibiotics were stopped, and who had an uneventful follow‐up of one year. With this new technology, one should be aware that DNA may persist after successful therapy, and therefore the presence of DNA does not always indicate persistent infection [54].

4  Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques

61

Figure 4.5.  Diagnostic algorithm in case of suspected BJI. Conventional culture methods should always be performed in order to diagnose and test the susceptibility of the pathogen(s). In addition to the conventional methods, non‐culture techniques are useful in special situations, as shown in the algorithm.

Key Points (Figure 4.5) ●●

●●

●●

●●

●●

Clinical samples should always be obtained for culture in order to detect and test the susceptibility of the pathogen(s). The use of non‐culture techniques may increase the rate of etiological diagnosis in pediatric BJI, particularly for fastidious microorganisms such as K. kingae. The use of a targeted PCR for S. aureus may shorten the time to diagnosis in cases of acute septic arthritis. Non‐culture techniques are a valuable supplemental tool in patients with culture‐ negative BJI caused by fastidious or slow‐growing microorganisms, allowing earlier initiation of pathogen‐adapted therapy. Non‐culture techniques as a supplemental tool improve the detection rate of the pathogen in patients with BJI who have received previous antibiotics.

References 1. Trampuz A, Piper KE, Jacobson MJ, et al. Sonication of removed hip and knee prostheses for diagnosis of infection. N Engl J Med 2007; 357:654–663. 2. Berbari EF, Marculescu C, Sia I, et al. Culture‐negative prosthetic joint infection. Clin Infect Dis 2007; 45:1113–1119.

62

Bone and Joint Infections

3. Bellova P, Knop‐Hammad V, Königshausen M, et al. Sonication of retrieved implants improves sensitivity in the diagnosis of periprosthetic joint infection. BMC Musculoskelet Disord 2019; 20: 623. 4. Baron EJ, Miller JM, Weinstein MP, et al. Executive summary: a guide to utilization of the microbiology laboratory for diagnosis of infectious diseases: 2013 recommendations by the Infectious Diseases Society of America (IDSA) and the American Society for Microbiology (ASM)(a). Clin Infect Dis 2013; 57:485–488. 5. Corvec S, Portillo ME, Pasticci BM, et al. Epidemiology and new developments in the diagnosis of prosthetic joint infection. Int J Artif Organs 2012; 35:923–934. 6. Roux AL, Sivadon‐Tardy V, Bauer T, et al. Diagnosis of prosthetic joint infection by beadmill processing of a periprosthetic specimen. Clin Microbiol Infect 2011; 17:447–450. 7. Portillo ME, Salvadó M, Alier A, et al. Advantages of sonication fluid culture for the diagnosis of prosthetic joint infection. J Infect 2014; 69:35–41. 8. Ascione T, Barrack R, Benito N, et  al. General Assembly, Diagnosis, Pathogen Isolation ‐ Culture Matters: Proceedings of International Consensus on Orthopedic Infections. J. Arthroplasty. 2019; 34:S197–S206. 9. Corvec S, Portillo ME, Vossen JA, et al. Diagnostics. In: AOTrauma Principles of Orthopedic Infection Management. 2016: 91–119. 10. Costerton JW, Post JC, Ehrlich GD, et al. New methods for the detection of orthopedic and other biofilm infections. FEMS Immunol Med Microbiol 2011; 61:133–140. 11. Gomez E, Cazanave C, Cunningham SA, et al. Prosthetic joint infection diagnosis using broad‐ range PCR of biofilms dislodged from knee and hip arthroplasty surfaces using sonication. J Clin Microbiol 2012; 50:3501–3508. 12. Portillo ME, Salvadó M, Trampuz A, et al. Sonication versus vortexing of implants for diagnosis of prosthetic joint infection. J Clin Microbiol 2013; 51:591–594. 13. Plouzeau C, Bémer P, Valentin AS, et  al. First experience of a multicenter external quality assessment of molecular 16S rRNA gene detection in bone and joint infections. J Clin Microbiol 2015; 53:419–424. 14. Reller LB, Weinstein MP, Petti CA. Detection and identification of microorganisms by gene amplification and sequencing. Clin Infect Dis 2007; 44:1108–1114. 15. Fenollar F, Roux V, Stein A, et al. Analysis of 525 samples to determine the usefulness of PCR amplification and sequencing of the 16S rRNA gene for diagnosis of bone and joint infections. J Clin Microbiol 2006; 44:1018–1028. 16. Kommedal Ø, Kvello K, Skjåstad R, et  al. Direct 16S rRNA gene sequencing from clinical specimens, with special focus on polybacterial samples and interpretation of mixed DNA chromatograms. J Clin Microbiol 2009; 47:3562–3568. 17. Bemer P, Plouzeau C, Tande D, et al. Evaluation of 16S rDNA PCR sensitivity and specificity for diagnosis of prosthetic‐joint infection: a prospective multicenter cross‐sectional study. J Clin Microbiol 2014; 52:3583–3589. 18. Jun Y, Jianghua L. Diagnosis of periprosthetic joint infection using polymerase chain reaction: an updated systematic review and meta‐analysis. Surg Infect (Larchmt) 2018; 19:555–565. 19. Hartley JC, Harris KA. Molecular techniques for diagnosing prosthetic joint infections – search results – PubMed. J Antimicrob Chemother 2014; 69:121–124. 20. Van Belkum A, Rochas O. Laboratory‐based and point‐of‐care testing for MSSA/MRSA detection in the age of whole genome sequencing. Front. Microbiol. 2018; 9:1437. 21. Ghebremedhin B, Layer F, Konig W, et  al. Genetic classification and distinguishing of Staphylococcus species based on different partial gap, 16S rRNA, hsp60, rpoB, sodA, and tuf gene sequences. J Clin Microbiol 2008; 46:1019–1025. 22. El Houmami N, Bzdreng J, Durand GA, et al. Molecular tests that target the RTX locus do not distinguish between Kingella kingae and the recently described Kingella negevensis Species. J Clin Microbiol 2017; 55:3113–3122. 23. Ross JJ. Septic Arthritis of Native Joints. Infect. Dis. Clin. North Am. 2017; 31:203–218.

4  Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques

63

24. Berbari EF, Kanj SS, Kowalski TJ, et al. Executive Summary: 2015 Infectious Diseases Society of America (IDSA) clinical practice guidelines for the diagnosis and treatment of native vertebral osteomyelitis in adults. Clin Infect Dis 2015; 61:859–863. 25. Liu K, Fu J, Yu B, et  al. Meta‐analysis of sonication prosthetic fluid PCR for diagnosing periprosthetic joint infection. PLoS One 2018; 13. 26. Portillo ME, Salvado M, Sorli L, et al. Multiplex PCR of sonication fluid accurately differentiates between prosthetic joint infection and aseptic failure. J Infect 2012; 65:541–548. 27. Malandain D, Bémer P, Leroy AG, et al. Assessment of the automated multiplex‐PCR Unyvero i60 ITI® cartridge system to diagnose prosthetic joint infection: a multicentre study. Clin Microbiol Infect 2018; 24:83.e1–83. 28. Vasoo S, Cunningham SA, Greenwood‐Quaintance KE, et al. Evaluation of the FilmArray blood culture ID panel on biofilms dislodged from explanted arthroplasties for prosthetic joint infection diagnosis. J. Clin. Microbiol. 2015; 53:2790–2792. 29. Dekker JP. Metagenomics for clinical infectious disease diagnostics steps closer to reality. J. Clin. Microbiol. 2018; 56:9. 30. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et al. Impact of contaminating DNA in whole‐genome amplification kits used for metagenomic shotgun sequencing for infection diagnosis. J Clin Microbiol 2017; 55:1789–1801. 31. Street TL, Sanderson ND, Atkins BL, et al. Molecular diagnosis of orthopedic‐device‐related infection directly from sonication fluid by metagenomic sequencing. J Clin Microbiol 2017; 55:2334–2347. 32. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et  al. Identification of prosthetic joint infection pathogens using a shotgun metagenomics approach ‐ PubMed. Clin Infect Dis 2018; 67:1333–1338. 33. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et al. A novel prosthetic joint infection pathogen, Mycoplasma Salivarium, identified by metagenomic shotgun sequencing – PubMed. Clin Infect Dis 2017; 65:332–335. 34. Tarabichi M, Alvand A, Shohat N,et al. Diagnosis of Streptococcus canis periprosthetic joint infection: the utility of next‐generation sequencing. Arthroplast Today 2018; 4:20–23. 35. Pham TT, Lazarevic V, Gaia N, et  al. Second periprosthetic joint infection caused by Streptococcus dysgalactiae: how genomic sequencing can help defining the best therapeutic strategy. Front Med 2020; 7:53. 36. Huang Z, Zhang C, Li W, et al. Metagenomic next‐generation sequencing contribution in identifying prosthetic joint infection due to Parvimonas micra: a case report. J Bone Jt Infect 2019; 4:50–55. 37. Sharma H, Ong MR, Ready D, et al. Real‐time whole genome sequencing to control a Streptococcus pyogenes outbreak at a national orthopaedic hospital. J Hosp Infect 2019; 103:21–26. 38. Ivy MI, Thoendel MJ, Jeraldo PR, et al. Direct detection and identification of prosthetic joint infection pathogens in synovial fluid by metagenomic shotgun sequencing. J Clin Microbiol 2018; 56:9. 39. Sanderson ND, Street TL, Foster D, et al. Real‐time analysis of nanopore‐based metagenomic sequencing from infected orthopaedic devices. Biological Sciences Genetics. Microbiology. BMC Genomics 2018; 19:714. 40. Chen MF, Chang CH, Chiang‐Ni C, et al. Rapid analysis of bacterial composition in prosthetic joint infection by 16s rRnA metagenomic sequencing. Bone Jt Res 2019; 8:367–377. 41. Li M, Zeng Y, Wu Y,et al. Performance of Sequencing Assays in Diagnosis of Prosthetic Joint Infection: A Systematic Review and Meta‐Analysis. J. Arthroplasty. 2019; 34:1514–1522. 42. Torchia MT, Austin DC, Kunkel ST, et al. Next‐generation sequencing vs culture‐based methods for diagnosing periprosthetic joint infection after total knee arthroplasty: a cost‐effectiveness analysis. J Arthroplasty 2019; 34:1333–1341. 43. Wang C, Huang Z, Li W, et al. Can metagenomic next‐generation sequencing identify the pathogens responsible for culture‐negative prosthetic joint infection? BMC Infect Dis 2020; 20:253.

64

Bone and Joint Infections

44. Deroche, Bémer, Valentin, et al. The right time to safely re‐evaluate empirical antimicrobial treatment of hip or knee prosthetic joint infections. J Clin Med 2019; 8:2113. 45. Carbonnelle E, Beretti JL, Cottyn S, et  al. Rapid identification of staphylococci isolated in clinical microbiology laboratories by matrix‐assisted laser desorption ionization‐time of flight mass spectrometry. J Clin Microbiol 2007; 45:2156–2161. 46. Kaleta EJ, Clark AE, Cherkaoui A, et al. Comparative analysis of PCR ‐ Electrospray ionization/mass spectrometry (MS) and MALDI‐TOF/MS for the identification of bacteria and yeast from positive blood culture bottles. Clin Chem 2011; 57:1057–1067. 47. Thomin J, Aubin GG, Foubert F, et al. Assessment of four protocols for rapid bacterial identification from positive blood culture pellets by matrix‐assisted laser desorption ionization‐time of flight mass spectrometry (Vitek® MS). J Microbiol Methods 2015; 115:54–56. 48. Harris LG, El‐Bouri K, Johnston S, et al. Rapid identification of staphylococci from prosthetic joint infections using MALDI‐TOF mass‐spectrometry. Int J Artif Organs 2010; 33:568–574. 49. Peel TN, Cole NC, Dylla BL, et al. Matrix‐assisted laser desorption ionization time of flight mass spectrometry and diagnostic testing for prosthetic joint infection in the clinical microbiology laboratory. Diagn Microbiol Infect Dis 2014; 81: 163–168. 50. Nagy E, Urbán E, Becker S, et al. MALDI‐TOF MS fingerprinting facilitates rapid discrimination of phylotypes I, II and III of Propionibacterium acnes. Anaerobe 2013; 20:20–26. 51. Lallemand E, Coiffier G, Arvieux C, et al. MALDI‐TOF MS performance compared to direct examination, culture, and 16S rDNA PCR for the rapid diagnosis of bone and joint infections. Eur J Clin Microbiol Infect Dis 2016; 35:857–866. 52. Greenwood‐Quaintance KE, Uhl JR, Hanssen AD, et al. Diagnosis of prosthetic joint infection by use of PCR‐electrospray ionization mass spectrometry. J Clin Microbiol 2014; 52:642–649. 53. Jacovides CL, Kreft R, Adeli B, et al. Successful identification of pathogens by polymerase chain reaction (PCR)‐based electron spray ionization time‐of‐flight mass spectrometry (ESI‐ TOF‐MS) in culture‐negative periprosthetic joint infection. J Bone Jt Surg ‐ Ser A 2012; 94:2247–2254. 54. Melendez DP, Uhl JR, Greenwood‐Quaintance KE, et al. Detection of prosthetic joint infection by use of PCR‐electrospray ionization mass spectrometry applied to synovial fluid. J Clin Microbiol 2014; 52:2202–2205. 55. Strålin K, Rothman RE, Özenci V, et  al. Performance of PCR/electrospray ionization‐mass spectrometry on whole blood for detection of bloodstream microorganisms in patients with suspected sepsis. J Clin Microbiol 2020. 56. Bémer P, Plouzeau C, Tande D, et al. Evaluation of 16S rRNA gene PCR sensitivity and specificity for diagnosis of prosthetic joint infection: A prospective multicenter cross‐sectional study. J Clin Microbiol 2014; 52:3583–3589.

Chapter 5

Bacteriophages for Treatment of Biofilm Infections Mercedes Gonzalez‐Moreno, Paula Morovic, Tamta Tkhilaishvili, and Andrej Trampuz

History of Bacteriophage Use in Human Infections Bacteriophages  –  or literally “bacteria‐eaters,” from the Greek phagein meaning “to devour” – were first mentioned in 1915 by Frederick Twort. They owe their therapeutic applications to Félix d’Hérelle, who isolated them in 1917 in stool samples of patients suffering from shigellosis [1]. Shortly thereafter, d’Herelle used bacteriophages to treat bacillary dysentery (shigellosis). This was probably the first attempt of phage application to treat pathogenic bacterial infections. The bacteriophage preparation was first ingested by d’Herelle and some colleagues in order to evaluate its safety before being administered to a 12‐year‐old boy with severe dysentery. After a single application of the anti‐­dysentery bacteriophage, the patient’s symptoms terminated, and the boy fully recovered within days [2]. Inspired by these results, d’Herelle continued studies on the therapeutic use of bacteriophages, carried many non‐randomized trials in humans [3], and co‐founded with George Eliava an institute, known today as the “Eliava Institute of Bacteriophages, Microbiology and Virology” in Tbilisi, Georgia, to carry out basic bacteriophage research and provide bacteriophages to treat human bacterial infections. The development of bacteriophages as antimicrobials continued for about three decades, i.e. from about 1915–1942 [4]. During this period, bacteriophages were used, among other indications, in France against avian typhoid caused by Salmonella gallinarum, and in the United States against chronic furunculosis. Phage therapy was also used during the Winter War between the former Soviet Union and Finland (1939–1940), with 6,000 Soviet soldiers treated against streptococcal or staphylococcal wound infections, which prevented limb amputations and reduced mortality due to gangrene. Companies such as Behring in Germany and Eli Lilly in the United States produced phage preparations against streptococci, staphylococci, and Escherichia coli. During World War II in Africa, the German army and the allied forces applied bacteriophages against dysentery [5].

Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

65

66

Bone and Joint Infections

The progress made on the synthesis of penicillin during the 1940s ushered in the era of antibiotic use, a golden age of medicine that largely continues to this day, which led to an almost complete abandonment of interest in the development of phages as clinically used antibacterial agents, especially in the Western countries [4]. In the Eastern countries, however, phage therapy was never abandoned, persisting to this day in countries such as Poland, Georgia, and Russia. Much knowledge of phage therapy in human patients comes from numerous publications in either Russian or Polish journals from that period, for example involving the use of oral phages to treat gastrointestinal infections, including shigellosis and salmonellosis [4]. Phage therapy was rediscovered by the English‐language literature starting with the work of Smith and Huggins in the 1980s, and progressively gained attention during the 1990s, followed by the start of human experiments in the 2000s [6]. The first placebo‐ controlled phase I trial in the United States was published in 2009, and showed no safety concerns [7]. The emergence of multidrug‐resistant bacterial infections has led to recent efforts investigating and promoting phage therapy to treat a multitude of infections. Despite the costly and time‐consuming requirements for the production of bacteriophages under current guidelines in the United States and the European Union, some countries are trying to accelerate the implementation of phage therapy through the so‐called “Magistral Approach.” Belgium, for instance, is currently implementing a pragmatic framework on phage therapy that centers on magistral preparation of individual therapeutic bacteriophages by pharmacies, and although the final products will not fully comply with the European requirements for medicinal products for human use (Directive 2001/83/EC), such magistral phage preparations can be used to treat patients in Belgium [8].

Principles of Bacteriophage Therapy Molecular Background Bacteriophages (also referred as phages) are viruses that specifically infect bacteria. Phages are the most abundant organisms on Earth, with an estimation of about 1031 particles distributed over all ecosystems on our planet. A phage is usually conformed by its genome (single or double‐stranded DNA or RNA) encapsulated in a protein capsid, which is sometimes completed with a tail and more or less complex appendages (e.g. spikes, tail fibers, etc.) (Figure 5.1) [9, 10]. As nonliving microorganisms, they rely on the bacterial cellular machinery to reproduce. The viral infection begins by attachment of the phage to its bacterial host through specific recognition of one or more receptors on the bacterial cell. These receptors can be found in the cell wall, bacterial capsules, slime layers, pili, or flagella, often consisting of proteins, lipopolysaccharides, teichoic acids, and other cell surface structures serving as irreversible phage‐binding receptors [11]. Upon recognition of the cell receptors, the phage injects its genetic material into the cytoplasm of the infected cell, and depending on its nature (virulent or temperate), it follows the lytic or lysogenic cycle. Virulent phages follow the lytic cycle, where the host’s genome is first degraded and the bacterial metabolic machinery is employed to copy the viral genome and produce viral proteins. After that, the viral particles are self‐assembled, and the bacterial cell is lysed by

5  Bacteriophages for Treatment of Biofilm Infections

67

(a) Myoviridae

Capsid

Collar Whiskers Tail tube and sheath Tail

Long tail fibre Short tail fibre Central tail fibre or spike

Baseplate

(b) Siphoviridae

(c) Podoviridae Capsid

Capsid Tail

Tail fibre Central tail fibre or spike

Tail tube Tail Baseplate

Tail fibre Central tail fibre or spike

Figure 5.1.  Representative structures of tailed phages. All tailed phages have a capsid that encloses and protects the genome and connects to the tail. (a) Phages in the Myoviridae family are the only tailed phages with a contractile tail sheath. (b) Both phages belonging to the Myoviridae and Siphoviridae families have a baseplate at the distal end of the tail to which receptor‐binding proteins (RBPs), such as tail fibers and tail spikes, are attached. (c) Because members of the Podoviridae have no baseplate, the RBPs directly attach to the tail. Siphoviridae and Podoviridae additionally have a central tail fiber or spike that protrudes from the distal end of the tail or baseplate. Reprinted with permission from Nobrega et al. [10].

68

Bone and Joint Infections

phage enzymes, releasing the progeny phages and killing the bacterial host. Temperate phages, on the other hand, can follow the lysogenic cycle. They become latent by inserting their genome either as a free plasmid inside the host cell or integrated into the bacterial chromosome. By this mean, they propagate to the next generations of bacterial cells. Under specific stressful environmental conditions, temperate phages can eventually shift towards the lytic cycle. As a result, the phage genome will be excised from the host ­chromosome, replicated, encapsulated, and then the phage particles will be released from the host bacterium by cell lysis, causing the death of the bacterial host cell [12]. Temperate phages have been described to transfer new genes to their hosts, including antibiotic resistance genes. In addition, they can alter the expression of host genes or provide ­protection to the host against infection by other phages [13]. Thus, strictly virulent phages, with immediate bactericidal effect, are favored for use in clinical practice.

Resistance Mechanisms Bacteria can develop resistance to phages at any stage of the phage infection cycle, classified as adaptive or non‐adaptive mechanism [11]. Common non‐specific anti‐phage resistance mechanisms that have been described until now include: (a) preclusion of phage adsorption and prevention of nucleic acid entry by surface modifications and receptor mutations; (b) superinfection exclusion systems preventing a secondary viral infection with the same or a closely related virus; (c) restriction‐modification systems responsible for the cleavage of exogenous dsDNA and protection of bacterial genetic material; and (d) abortive infection leading to cell death or stasis when phage replication takes place. A second line of defense (adaptive defense) is associated with restriction enzymes and the CRISPR/Cas system (Clustered Regularly Interspaced Short Palindromic Repeats and associated proteins), which can identify and cleave phage nucleic acids in a highly specific and effective manner [14]. Phages can evolve and develop counterstrategies to circumvent bacterial anti‐phage mechanisms. Based on their genomic plasticity and rapid replication rates, phages can overcome adsorption inhibition by point mutations in specific genes or escape from restriction‐modification mechanisms by genome rearrangements. Furthermore, phages can use anti‐CRISPR proteins to evade the CRISPR/Cas system, or avoid the abortive infection by hijacking bacterial antitoxins [11]. Contrary to antibiotics, phages have minimal influence on the normal microbiome due to their high bacterial species or strain specificity, and have the ability to increase in number at the site of infection due to their “multiplicity” [15], which theoretically would imply that a little phage dose is sufficient for effective treatment.

Antibiotic–Bacteriophage Interactions Although the use of phages alone would potentially demonstrate clinical success, their combination with antibiotics have shown to be more effective than phage monotherapy in numerous in vitro and animal studies. These studies have proved statistically or clinically significant phage‐antibiotic synergism, biofilm minimization, or reductions in resistance emergence [12]. Regardless of the antibiotic resistance state of the bacteria, the combinatorial approach using phages and antibiotics has demonstrated a range of benefits [16]. For instance, it has been shown for some phage/antibiotic combinations that sub‐inhibitory

5  Bacteriophages for Treatment of Biofilm Infections

69

concentration of antibiotics can foster phage productivity and consequently decrease ­bacterial counts. A restoration of antibiotic sensitivity, by loss of bacterial fitness or by phage interaction with the bacterial drug efflux systems, has also been described. Phages can additionally work as adjuvants of antibiotics against biofilms by enabling antibiotics to reach bacterial cells deeper within biofilms through degradation of the exopolysaccharide matrix by depolymerases and by infecting antibiotic‐tolerant persisters (further information in section “Activity of Phages against Bacterial Biofilms and Persisters”). However, antagonistic effects can also occur with phage/antibiotic combinations depending on the treatment conditions (e.g., dosage, order of administration, timing, etc.). Thus, combinatorial therapies require a careful choice of dosing and time points at which either antimicrobial substance is administered. In vitro studies determined better outcomes if phage was administered before antibiotics than if antibiotics were introduced before or simultaneously with phage. This is possibly due to the killing of host bacteria, which are essential for phage production, by antibiotics [12]. Other competing dynamics between phage and antibiotics may also play a role. Antibiotics are expected to interfere with aspects of bacterial physiology that can be crucial to phage activities, as for instance by interfering with bacterial ribosome functioning necessary for phage protein production [17]. The pharmacokinetic/pharmacodynamic (PK/PD) modeling techniques traditionally used for antibiotics differ for phages. The phage concentration is expected to increase at the site of infection through their replication in living bacteria. PK/PD models for phage therapy should integrate the classical antimicrobial pharmacological view (drug impact on the body, drug interactions, absorption, distribution, metabolism, secretion, etc.) with the self‐replication characteristic of phages [18]. The immune system plays also a key role in phage inactivation and/or clearance from the body, which may pose a problem maintaining sufficient phage titers for therapeutic activity. Based on limited reports on immune response in clinical studies using virulent phages, their immunogenicity does not seem to represent a safety risk. Major concerns encompass an increase in pro‐inflammatory cytokines as response to the potentially massive liberation of bacterial endotoxins after bacterial lysis, as has been observed with the use of certain antibiotics [19]. So far, there is not enough evidence‐based data for a better understanding of the phage pharmacokinetics and the phage immune interaction as well as the clinical relevance of all these parameters.

Activity of Bacteriophages against Bacterial Biofilms and Persisters Biofilms are complex clusters of bacteria, formed by single or multiple species, merged by extracellular polymeric substances (EPS) and adhered to surfaces, including living tissue or medical devices, among others. Biofilm microorganisms are metabolically less active and have a minimal growth rate. Therefore, they are tolerant to many antibiotics [20][21]. Bacteriophages showed promising results for biofilm eradication due to their multiplicity at the infection site, but also by producing specific enzymes that allow them to actively penetrate and disrupt biofilms, and for their ability to infect persisters, which are less metabolically active bacteria [20]. Phages encoding EPS‐degrading enzymes are particularly useful against biofilms. A diverse group of phage‐encoded enzymes, called depolymerases, capable of degrading polymers – either associated with the cell surface (e.g. capsule polysaccharides) to facilitate phage adsorption, or EPS involved in biofilm matrix in order to promote phage d ­ iffusion

70

Bone and Joint Infections

through the biofilm  –  have been described [22]. Depolymerases can be associated with virions, forming part of the phage particle (e.g. in their tail spikes), or be in soluble form. Depolymerases derived from phages have been tested against biofilms of different bacterial species, exhibiting dose‐dependent activity and reducing significantly the biofilm biomass [20, 23, 24]. Similar to the host specificity of bacteriophages, phage‐associated depolymerases can be highly specific for host‐derived EPS. Since different species of bacteria produce different EPS components, depolymerase active against the polysaccharides produced by one species may not act on that produced by other bacteria [24]. However, some depolymerases are capable of degrading EPS of several genera [23]. Moreover, some bacteriophages can induce their host bacteria to produce and release depolymerases, which could be a phage mechanism to make the biofilm matrix more porous, facilitating infection by progeny bacteriophage or, alternatively, a fight response by infected bacteria, seeking to facilitate movement away from the focus of infection [24], leading in any case to a disaggregation of the biofilm. Unlike other antimicrobials, bacteriophages replicate within their host cells, which can result in self‐sustaining infections with ongoing amplification leading to an increasing number of bacteriophages. The localized spread of phage progeny continues infecting and killing more bacteria, which is called multiplicity at the infection site. These mechanisms require a critical mass of host bacteria at the same location, which is typically the case in biofilm infections [24]. Hence, by spreading through the biofilm, bacteriophages can progressively remove the biofilm and reduce the potential for regrowth. The regrowth of bacteria within the biofilm is thought to arise from the presence of persisters. Unlike resistant bacterial cells, where resistance mechanisms are based on genetic changes that block antimicrobial activity, persisters present a transient non‐heritable phenotype that is thought to be less sensitive to antibiotics because the cells are not undergoing cellular activities that antibiotics can corrupt, which results in tolerance [25]. Thus, persisters can remain viable over the course of antibiotic exposure and repopulate the biofilm when the levels of antibiotic drop, causing the relapse of the infection. Some studies have reported on the ability of bacteriophages to infect persisters, and initiate a productive lytic infection when persisters switch to normal growth, ultimately causing their lysis [24]. New concepts are emerging nowadays in the design of phage‐based treatments to maximize phage therapy efficacy and minimize the likelihood of resistance emergence. A schematic illustration of phage‐based treatments for biofilm removal is shown in Figure 5.2 [20].

Phage-based treatments for biofilm removal

Phage therapy

Single phages

Genetically modified phages

Phage-derived enzymes

Phage + Antibiotics

Phage cocktails

Figure 5.2.  Main phage‐based treatments for biofilm removal. Reprinted with permission from Ferriol‐González and Domingo‐Calap [20].

5  Bacteriophages for Treatment of Biofilm Infections

71

Designing phage cocktails that include phages against multiple species has been shown to be especially effective against multi‐species biofilms [23, 26]. Phage cocktails, besides conferring activity against a broader host range, can help also in preventing the emergence of phage‐resistant bacteria if multiple phages active against a given target are included in the cocktail [20]. However, in order to avoid possible undesired effects when using phage cocktails, a rational approach to designing cocktails is crucial. In a phage cocktail, the various phages should not compete with each other, in order to minimize the risk for reduction of efficacy. In addition, the mechanisms of phage resistance by bacteria should be different in order to minimize the risk of cross‐resistance [26]. Bacteriophages can be genetically modified to improve their bacterial killing properties. Existing examples of genetically engineered phages include a phage with altered tail fiber proteins to extend its host range [23], a phage designed to produce a soluble hydrolase that enhances biofilm degradation, a temperate phage turned into a lytic phage by removal of all genes related to lysogeny or a chimeric phage encoding a short peptide with broad‐spectrum anti‐biofilm effect [20].

Bacteriophage Susceptibility Testing: the Phagogram Phage therapy is still not approved in most parts of the world, and extensive research on its efficacy and safety is still to be conducted. However, if the particular conditions according to the Declaration of Helsinki (article 37) are met and bacteriophages are to be applied, the magistral approach (“compounded” drug product in United States) is to be followed. In practice, it means that, only when no other option of treatment is available, a phage preparation is prescribed by a physician for an individual patient and prepared by the hospital pharmacist following strict safety regulations [8]. Due to the high host‐specificity of bacteriophages (mostly infecting single species or even single strains of bacteria), when preparing the phage solution for an individual patient, it is important to select bacteriophages active against the patient’s isolated strain. To this end, similarly to an antibiogram (antibiotic susceptibility testing), a so‐called “phagogram” needs to be performed [8]. Various methods for testing bacteria susceptibility to bacteriophages have been described, such as the spot test, efficacy of plating (EOP), or killing assays. The simplest method among them is the spot test, where small droplets of a bacteriophage lysate are applied on a plate prepared with the bacterial strain to be tested and the appearance of a clear zone (lysis) determines bacterial susceptibility to the bacteriophage. However, lysis observed by this method may be the result not only from the bacteriophage infection that gives rise to lysis and production of new phage, but also due to residues of bacteriocins on the phage lysate that kills bacteria, or to the phages themselves causing abortive infections or lysis from without, leading to false‐positive results [27]. Performing an EOP assay, by plating different titers of bacteriophages, quantifying plaque‐forming units (PFU), and comparing it to the PFU count of a reference bacterial strain can provide more information on the efficacy of a particular bacteriophage [27]. Nevertheless, the absence of plaque formation does not necessarily correlate with a lack of bacteriophage ability for a productive infection. Plaque formation might depend on several factors, including phage diffusion in agar, adsorption rate, electrolyte requirements, growth phase of the host, etc. [28]. Killing assays, in which bacteria and bacteriophages are incubated together in liquid medium and the optical density or heat flow production are measured as indicators of bacterial presence, represent useful methods for determination of the minimum phage

72

Bone and Joint Infections

titer needed for successful bacteria killing or to better monitor phage virulence [29, 30]. On the other hand, these assays are usually less cost‐efficient than clinical laboratory tests due to a higher instrumental cost, minimum automation, or limitations in their throughput. A standardized method that is easy to perform, fast, and available to everyone is still to be developed. Currently, several projects for the development of an automated, reliable, and reproducible phagogram technique are ongoing. Some examples include the PHAGOGramme project under development by Pherecydes Pharma (https://www.pherecydes‐pharma. com/phagogramme.html), the combined PhageBank™and HRQT™approaches of the clinical‐stage company Adaptive Phage Therapeutics (http://www.aphage.com/the‐ science/#phagebank) or the PhagoFlow project as a joint effort of different institutions including the Charité‐Universitätsmedizin, the Bundeswehr Hospital Berlin, the Leibniz Institute DSMZ, and the Fraunhofer ITEM (https://www.phagoflow.de/en/phagogram/).

Experimental and Clinical Evidence with Bacteriophage Treatment Despite the long history of use of phages for antibacterial therapy since their discovery in the early 1900s, and even with the availability of phage products for the treatment of bacterial infections in some countries (e.g. Georgia, Poland, Russia), extensive in vitro and experimental studies as well as clinical trials to fulfill the requirements for phage therapy according to good manufacturing practice guidelines are lacking [31]. The efficacy of phage therapy has been investigated for bloodstream, gastrointestinal, urinary tract, and respiratory infections, and burn wounds [19, 26]. We limit our focus on experimental and clinical evidence of phage therapy in bone and joint infections. Most preclinical studies investigated the efficacy of bacteriophages on monomicrobial S. aureus or P. aeruginosa infections demonstrating large reduction of planktonic bacteria, successful prevention of bacterial adherence to foreign material, and synergism between antibiotics and phages to eradicate biofilms [32]. However, numerous experimental limitations need to be addressed. For instance, limited data exist on phages active against S. epidermidis despite its high prevalence in implant‐associated infections, its strong biofilm‐forming capability, and its extensive resistance to antibiotics [33]. Moreover, in vivo models that replicate the joint and peri‐implant microenvironment are lacking, which makes the translation of preclinical findings into clinical settings difficult [34]. One promising in vivo model published by Carli et al. in 2017 [35] replicates accurately the clinical setting of total joint replacement, and could therefore be adopted in the future for phage therapy testing. Other studies reported a concentration dependency of phage therapy, suggesting that low‐titer phage administration or single instead of multiple doses are unlikely to be successful. In addition, considering, for instance, possible vascular impairments in open fractures or the wish for reduction of systemic effects, local treatment is often preferred in bone and joint infections [36]. Hence, in aiming for phage stability and appropriate release kinetics during treatment, an important part of research in phage therapy is focused on the encapsulation of phages into sustained release systems. Numerous strategies regarding bacteriophage formulation and encapsulation are being implemented, showing promising outcomes under experimental settings [36]. Still, great challenges involve a rational design of carriers loaded with precise doses of encapsulated phage able to support controlled releases in patients. Osteomyelitis is another clinical field where phage therapy has been applied, often using anti‐staphylococcal bacteriophages. A summary of clinical studies on phage

5  Bacteriophages for Treatment of Biofilm Infections

73

t­ herapy for musculoskeletal infections is presented in Table 5.1 [32]. The largest clinical study with 120 participants was conducted in Tbilisi, Georgia, assessing the therapeutic efficacy of a custom‐made staphylococcal cocktail against arthritis and osteomyelitis. The summarized results do not allow us to evaluate phage efficacy. All 120 patients had complete recovery of osteomyelitis and/or arthritis, namely 9 patients with phage therapy alone, 51 patients with phage plus antibiotics, and 60 patients with antibiotics alone. These results are much better in each treatment group than could be expected. In western countries, due to strict regulations in application of phage therapy, clinical experience with bacteriophages is limited to individual cases with a total of 5 case reports and 1 case series published between 2017 and 2019. As shown in Table  5.1, two case reports investigated the use of bacteriophages in osteomyelitis, another two in prosthetic Table 5.1.  Human clinical studies on phage therapy for musculoskeletal infections. Reprinted with ­permission from Onsea et al. [32]. Reference

Sample Patient size characteristics

Intervention

Outcome

Lang et al., 1979

7

PJI (n = 2), OM (n = 1), Septic arthritis (n = 1), Spinal infection (n = 1), FRI (n = 2).

Phages adapted to isolated strains. Administration either topical or by injection through a draining system. Some cases received combination treatment with antibiotics.

5/7 treated Recurrence of spinal infection and one FRI.

Kutateladze and Adamia, 2010

120

Patients with staphylococcal OM or arthritis.

Three groups: ‐ antibiotics (n = 60) ‐ phage monotherapy (n = 9) ‐ phage + antibiotics (n = 51). Administration of Eliava staphylococcal phage preparation topically or intravenously.

100% success rate in all groups.

Slopek et al., 1987

100

Purulent arthritis Administration locally and/or orally. Some cases received combination and myositis treatment with antibiotics. (n = 19), OM of the long bones (n = 40), FRI (n = 41).

Weber‐ Dabrowska et al., 2000

81

OM of the long bones (n = 40), FRI (n = 41).

Administration locally and/or orally. Unclear if some patients received combination treatment with antibiotics.

Vogt et al., 2017

1

OM

Repeated dosing of phage cocktail. Eradication of the Pyo bacteriophage through draining infection. system, in combination with antibiotic therapy.

Ferry et al., 2018a

1

OM (post‐radiation)

Application of customised phage cocktail every 3 d, in combination with intravenous antibiotic therapy.

Patient died 45 d after treatment due to cancer progression.

Ferry et al., 2018b

1

PJI

Single intraoperative injection of a customised phage cocktail in combination with intravenous antibiotic therapy.

Eradication of the infection.

Success rates: ‐ purulent arthritis and myositis: 89.5% ‐ OM of the long bones: 95% ‐ FRI: 90.2% Success rates: OM of the long bones: 95%FRI 60%

(Continued )

74

Bone and Joint Infections

Table 5.1.  (Continued) Reference

Sample Patient size characteristics

Intervention

Outcome

Nir‐Paz et al., 2019

1

FRI

Intravenous repeated administration of customised phage cocktail, in combination with intravenous antibiotic therapy.

Eradication of the infection (after two phage therapy regimens).

Tkhilaishvili et al., 2019

1

PJI

Repeated dosing of customised phage cocktail, in combination with intravenous antibiotic therapy.

Eradication of the infection.

Onsea et al., 2019

4

OM

Repeated dosing of BFC1 cocktail or Pyo bacteriophage cocktail in combination with intravenous antibiotic therapy.

Eradication of the infection in all cases.

Abbreviations: PJI, periprosthetic joint infection; OM, osteomyelitis; FRI, fracture‐related infection.

joint infection, and one in a fracture‐related infection. The applied bacteriophages were used to target P. aeruginosa, S. aureus, Acinetobacter baumannii, Klebsiella pneumoniae, S. epidermidis, and E. faecalis, respectively. Bacteriophages were administered either intravenously or locally, and in combination with intravenous antibiotics. Eradication of infection was seen in all case reports except for one, where the outcome of infection was unclear due to death caused by the primary disease of the patient. Although bacteriophages have so far demonstrated good efficacy and safety, experience is still lacking, and comprehensive and well‐organized studies on the production and processing of bacteriophages, their administration, and dosage, as well as exhaustive clinical monitoring of results, are still needed.

Local Delivery and Systemic Bacteriophage Application A major hurdle of phage therapy is the achievement of a sufficient number of phages at the site of infection to accomplish therapeutic activity. Phage ability to disseminate throughout the body strongly depends on the route of administration and the initial phage dose [37]. Administration of high or repeated phage doses might increase the chances for successful distribution. Furthermore, encapsulation of phages might allow a controlled phage release and act as a shield against chemical degradation or immunological neutralization, prolonging its systemic circulation period [18]. Routes of phage application include the use of parenteral administration, being oral dosing, topical, and aerosolization also commonly applied. A summary of some advantages and disadvantages of the administration routes can be seen in Table 5.2 ([19]).

Systemic Delivery Systemic phage delivery by intravenous, intraperitoneal, or intramuscular injection allows rapid phage dissemination in different organs and tissues such as the liver, spleen, kidneys, and lungs [38]. A nearly complete recovery of administered phages was shown

5  Bacteriophages for Treatment of Biofilm Infections

75

Table 5.2.  Routes of administration for phage therapy. Reprinted with permission from Romero‐Calle et al. [19]. Delivery Route

Advantages

Disadvantages

Mitigations to Hurdles

Intraperitoneal

Higher dosage volumes possible. Diffusion to other sites.

Extent of diffusion to other sites may be overestimated in humans (most data from small animals).

Multiple delivery sites.

Intramuscular

Phages delivered at infection site.

Slower diffusion of phages (possibly). Lower dosage volumes.

Multi‐dose courses.

Subcutaneous

Localized and systemic diffusion.

Lower dosage volumes.

Multi‐dose courses.

Intravenous

Rapid systemic diffusion.

Rapid clearing of phages by the immune system.

In vivo selection of low‐immunogenic phages may be possible.

Topical

High dose of phages delivered at infection site.

Run‐off from target site if phages suspended in liquid.

Incorporate phages into gels and dressings.

Suppository

Slow, stable release of phages over long time.

Limited applications/sites. Risk of insufficient dosing. Technically challenging to manufacture.

Careful consideration of phage kinetics required.

Oral

Ease of delivery. Higher dosage volumes possible.

Stomach acid reduces phage titer. Non‐specific adherence of phages to stomach contents and other microflora.

Add calcium carbonate to buffer pH. Microencapsulation to deliver phages to target area.

Aerosol

Relative ease of delivery. Can reach poorly perfused regions of infected lungs.

High proportion of phages lost. Delivery can be impaired by mucus and biofilms.

Use of depolymerases to reduce mucus.

several minutes after intravenous application [37]. Phage distribution was documented in the heart, skeletal muscles, bladder, thymus, bone narrow, lungs, and brain but not yet in joints, bone, or eyes [18]. The use of highly purified phage preparations without bacterial components or endotoxins is essential to minimize the risk of side effects due to impurities.

Oral or Inhaled Delivery Oral route of administration has been successfully used in gastrointestinal infections. However, phage stability in acidic environments in the stomach and duodenum may reduce the phage concentration or activity [38]. Thus, protection of phages from the gastric acidity could be achieved by phage encapsulation, as shown in a study against Salmonella spp. [39]. In respiratory infections, liquid and dry powder phage formulations were investigated for nebulization and inhalation for topical delivery in acute and chronic lung infections [40].

76

Bone and Joint Infections

Local Delivery Effective local delivery of antibacterial substances is essential in patients with biofilm infections associated either with implanted medical devices or chronic wounds, since antibacterial drugs have limited activity in such infections. Bacteriophages may have better efficacy in such infections, provided that they reach the infectious site. Therefore, a major focus is being set in the implementation of drug carriers to allow local and prolonged release of phages. Unfortunately, there is insufficient data available on processing phages into well‐defined pharmaceutical formulations, their long‐term stability, and impact on phage efficacy in vivo [40]. Wound healing is one of the therapeutic areas where local application of phages has received a lot of attention. Advances in the development of phage formulations including hydrogels, liposome entrapment, or phage‐immobilized wound dressings have led to increasing successful rates in the topical application of phage therapy [41]. Numerous reviews report on the widespread clinical use of phage preparations for the treatment of skin infections, and purulent and surgical wounds, mostly by the former Soviet Union countries. In Europe, the project Phagoburn, launched in June 2013, was the first prospective multicentric, randomized, single‐blind, and controlled clinical trial on phage therapy to treat Escherichia coli and Pseudomonas aeruginosa skin infections in burn patients [42]. It allowed significant advances regarding the regulatory framework of phage therapy as well. Some clinical cases and preclinical studies also support the effective local delivery of bacteriophages to treat local bone infections (see section “Experimental and Clinical Evidence with Bacteriophage Treatment”). Currently, the application of phages in patients with severe musculoskeletal infections has generally consisted in local administration through a draining system (Figure 5.3) [43]. Although this approach has shown successful outcomes, it has the drawback of the usage of the drainage tube as delivery route, which could favor the emergence of superinfections, besides being a cumbersome method. Thus, the optimization of local phage delivery strategies might help in overcoming these issues. Some examples are an engineered hydrogel for controlled delivery of phage targeting Pseudomonas aeruginosa to the site of orthopedic infections [44] or the use of fibrin glue for sustained delivery of viable phages [45]. Phages have also been immobilized on surfaces for the prevention of biofilm formation with examples on urinary catheters or on nylon sutures for wound healing applications [36].

Outlook and Future Perspectives The rising threat of multiresistant bacterial infections has brought together many research institutions, hospitals, and the industry in a joint effort to seek alternative treatments to the conventional use of antibiotics. Phages have unique features that make them convincing antibacterial agents, alone or in combination with other antimicrobials, while the constraints associated with the implementation of phage therapy could be overcome through a combination of proper phage selection, effective formulation, and greater clinician understanding of and familiarity with product application. Phages have been used to treat bacterial infections since their discovery, being for decades and also today the standard of care in several countries of Eastern Europe and having demonstrated clinical success in recent compassionate care cases in Western Europe and the United States, with no serious adverse events reported to date.

5  Bacteriophages for Treatment of Biofilm Infections

(a)

(b)

(c)

(d)

77

Figure 5.3.  Phage therapy of pelvic osteomyelitis (pathogen: pan‐resistant Pseudomonas aeruginosa). (a) Removal of foreign bodies and surgical debridement of necrotic tissue. (b) Preparation of a wound filler impregnated with the phage solution. (c) Insertion of the instillation tubes before wound closure. (d) Daily administration of 50 ml of the bacteriophage suspension after pre‐treatment of the wound with bicarbonate buffer (over one week). Reprinted with permission from Vogt et al. [43].

The increasing number of publications that have appeared during the last decade and the growing interest of the industry in phage therapy represent very encouraging progress in addressing the knowledge gap required for phage therapeutic applications.

Key Points ●●

With increasing antimicrobial resistance of bacteria, there is a rising interest in the therapeutic potency of bacteriophages.

78

●●

●●

●●

Bone and Joint Infections

Biofilm infections are tolerant to most antibiotics. Phage therapy of such infections could be an attractive new option. There are a few clinical studies showing that bacteriophages are able to eradicate musculoskeletal infections without serious adverse events. However, controlled trials of high quality are still lacking. Despite current restrictions in the application of phage therapy, commercialization of phage‐based technologies is gaining interest in Western countries.

References 1. D’Herelle F. On an invisible microbe antagonistic toward dysenteric bacilli: brief note by Mr. F. D’Herelle, presented by Mr. Roux. 1917. Res Microbiol, 2007. 158(7):553–554. 2. Sulakvelidze A, Alavidze Z, Morris JG, Jr. Bacteriophage therapy. Antimicrob Agents Chemother, 2001. 45(3):649–659. 3. Dublanchet A and Bourne S. The epic of phage therapy. Can J Infect Dis Med Microbiol, 2007. 18(1):15–18. 4. Green S, Ma L, Maresso A. (2019). Phage therapy, in T.M. Schmidt (ed.) Encyclopedia of Microbiology (Fourth Edition), pp. 485–495 (Oxford: Academic Press). 5. Moelling K, Broecker F, Willy C. A wake‐up call: We need phage therapy now. Viruses, 2018. 10(12). doi: 10.3390/v10120688. 6. Wittebole X, De Roock S, Opal SM. A historical overview of bacteriophage therapy as an alternative to antibiotics for the treatment of bacterial pathogens. Virulence, 2014. 5(1):226–235. 7. Rhoads DD, Wolcott RD, Kuskowski MA, et al. Bacteriophage therapy of venous leg ulcers in humans: results of a phase I safety trial. J Wound Care, 2009. 18(6):237–243. 8. Pirnay JP, Verbeken G, Ceyssens PJ, et  al. The magistral phage. Viruses, 2018. 10(2). doi: 10.3390/v10020064. 9. Dion MB, Oechslin F, Moineau S. Phage diversity, genomics and phylogeny. Nat Rev Microbiol, 2020. 18(3):125–138. 10. Nobrega FL, Vlot M, de Jonge PA, et  al. Targeting mechanisms of tailed bacteriophages. Nat Rev Microbiol, 2018. 16(12):760–773. 11. Orzechowska B and Mohammed M. (2019). The war between bacteria and bacteriophages, in M. Mishra (ed.) Growing and Handling of Bacterial Cultures, (London: IntechOpen). doi: 10.5772/intechopen.87247. 12. Morrisette T, Kebriaei R, Lev KL, et al. Bacteriophage therapeutics: A primer for clinicians on phage‐antibiotic combinations. Pharmacotherapy, 2020. 40(2):153–168. 13. Howard‐Varona C, Hargreaves KR, Abedon ST, et al. Lysogeny in nature: mechanisms, impact and ecology of temperate phages. ISME J, 2017. 11(7):1511–1520. 14. Ofir G and Sorek R. Contemporary phage biology: from classic models to new insights. Cell, 2018. 172(6):1260–1270. 15. Loc‐Carrillo C and Abedon ST. Pros and cons of phage therapy. Bacteriophage, 2011. 1(2):111–114. 16. Tagliaferri TL, Jansen M, Horz HP. Fighting pathogenic bacteria on two fronts: Phages and antibiotics as combined strategy. Front Cell Infect Microbiol, 2019. 9(22). doi: 10.3389/ fcimb.2019.00022. 17. Abedon ST. Phage‐antibiotic combination treatments: Antagonistic impacts of antibiotics on the pharmacodynamics of phage therapy? Antibiotics, 2019. 8(4). doi: 10.3390/antibiotics8040182. 18. Dąbrowska K. Phage therapy: What factors shape phage pharmacokinetics and bioavailability? Systematic and critical review. Med Res Rev, 2019. 39(5):2000–2025.

5  Bacteriophages for Treatment of Biofilm Infections

79

19. Romero‐Calle D, Guimarães‐Benevides R, Góes‐Neto A, et al. Bacteriophages as alternatives to antibiotics in clinical care. Antibiotics, 2019. 8(3). doi: 10.3390/antibiotics8030138. 20. Ferriol‐González C and Domingo‐Calap P. Phages for biofilm removal. Antibiotics, 2020. 9(5). doi: 10.3390/antibiotics9050268. 21. Stewart PS. Antimicrobial tolerance in biofilms. Microbiol Spectr, 2015. 3(3). doi: 10.1128/ microbiolspec.MB‐0010‐2014. 22. Pires DP, Oliveira H, Melo LDR, et al. Bacteriophage‐encoded depolymerases: their diversity and biotechnological applications. Appl Microbiol Biotechnol, 2016. 100(5):2141–2151. 23. Geredew‐Kifelew L, Mitchell JG, Speck P. Mini‐review: efficacy of lytic bacteriophages on multispecies biofilms. Biofouling, 2019. 35(4):472–481. 24. Harper DR, Parracho HMRT, Walker J, et al. Bacteriophages and biofilms. Antibiotics, 2014. 3(3):270–284. 25. Balaban NQ, Helaine S, Lewis K, et al. Definitions and guidelines for research on antibiotic persistence. Nat Rev Microbiol, 2019. 17(7):441–448. 26. Kortright KE, Chan BK, Koff JL, et al. Phage therapy: A renewed approach to combat antibiotic‐resistant bacteria. Cell Host Microbe, 2019. 25(2):219–232. 27. Khan‐Mirzaei M and Nilsson AS. Isolation of phages for phage therapy: A comparison of spot tests and efficiency of plating analyses for determination of host range and efficacy. PLoS One, 2015. 10(3). doi: 10.1371/journal.pone.0118557. 28. Abedon S. (2018). Detection of bacteriophages: phage plaques, in D. Harper, et  al. (ed.) Bacteriophages, pp. 1–32. (Basel: Springer, Cham). 29. Xie Y, Wahab L, Gill JJ. Development and validation of a microtiter plate‐based assay for determination of bacteriophage host range and virulence. Viruses, 2018. 10(4). doi: 10.3390/ v10040189. 30. Tkhilaishvili T, Di Luca M, Abbandonato G, et  al. Real‐time assessment of bacteriophage T3‐derived antimicrobial activity against planktonic and biofilm‐embedded Escherichia coli by isothermal microcalorimetry. Res Microbiol, 2018. 169(9):515–521. 31. Alavidze Z, Ami‐nov R, Betts A, et al. Silk route to the acceptance and re‐implementation of bacteriophage therapy. Biotechnol J, 2016. 11(5):595–600. 32. Onsea J, Wagemans J, Pirnay JP, et al. Bacteriophage therapy as a treatment strategy for orthopaedic‐device‐related infections: where do we stand? Eur Cell Mater, 2020. 39:193–210. 33. Akanda ZZ, Taha M, Abdelbary H. Current review—The rise of bacteriophage as a unique therapeutic platform in treating peri‐prosthetic joint infections. J Orthop Res, 2018. 36(4):1051–1060. 34. Jie K, Deng P, Cao H, et al. Prosthesis design of animal models of periprosthetic joint infection following total knee arthroplasty: A systematic review. PLoS One, 2019. 14(10). doi: 10.1371/ journal.pone.0223402. 35. Carli AV, Bhimani S, Yang X, et al. Quantification of peri‐implant bacterial load and in vivo biofilm formation in an innovative, clinically representative mouse model of periprosthetic joint infection. J Bone Joint Surg Am, 2017. 99(6). doi: 10.2106/jbjs.16.00815. 36. Malik DJ, Sokolov IJ, Vinner GK, et al. Formulation, stabilisation and encapsulation of bacteriophage for phage therapy. Adv Colloid Interface Sci, 2017. 249:100–133. 37. Huh H, Wong S, St. Jean J, et al. Bacteriophage interactions with mammalian tissue: Therapeutic applications. Adv Drug Deliv Rev, 2019. 145:4–17. 38. Ryan EM, Gorman SP, Donnelly RF, et al. Recent advances in bacteriophage therapy: how delivery routes, formulation, concentration and timing influence the success of phage therapy. J Pharm Pharmacol, 2011. 63(10):1253–1264. 39. Colom J, Cano‐Sarabia M, Otero J, et al. Liposome‐encapsulated bacteriophages for enhanced oral phage therapy against Salmonella spp. Appl Environ Microbiol, 2015. 81(14):4841–4849. 40. Chang RYK, Wallin M, Lin Y, et al. Phage therapy for respiratory infections. Adv Drug Deliv Rev, 2018. 133:76–86. 41. Chang RYK, Morales S, Okamoto Y, et al. Topical application of bacteriophages for treatment of wound infections. Transl Res, 2020. 220:153–166.

80

Bone and Joint Infections

42. Jault P, Leclerc T, Jennes S, et al. Efficacy and tolerability of a cocktail of bacteriophages to treat burn wounds infected by Pseudomonas aeruginosa (PhagoBurn): a randomised, controlled, double‐blind phase 1/2 trial. Lancet Infect Dis, 2019. 19(1):35–45. 43. Vogt D, Sperling S, Tkhilaishvili T, et al. Beyond antibiotic therapy“ – Zukünftige antiinfektiöse Strategien – Update 2017. Der Unfallchirurg, 2017. 120(7):573–584. 44. Wroe JA, Johnson CT, García AJ. Bacteriophage delivering hydrogels reduce biofilm formation in vitro and infection in vivo. J Biomed Mater Res A, 2020. 108(1):39–49. 45. Rubalskii E, Ruemke S, Salmoukas C, et al. Fibrin glue as a local druge‐delivery system for bacteriophage PA5. Sci Rep, 2019. 9(1). doi: 10.1038/s41598‐018‐38318‐4.

Chapter 6

Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone Cornelia B. Landersdorfer, Jürgen B. Bulitta, Roger L. Nation, and Fritz Sörgel

Chronic osteomyelitis requires prolonged antibiotic treatment, has a high recurrence rate, and can cause irreversible damage. Also, the number of orthopedic device‐related infections is projected to increase [1,2]. Therefore, adequate antibiotic treatment and surgical prophylaxis are critical. Therapeutic success is primarily determined by the antimicrobial activity against the infecting pathogen and the rate and extent of antibiotic penetration into bone. Adequate bone penetration has to be ensured as antibiotics need to reach effective concentrations at the infection site to kill bacterial pathogens. Therefore, studying the time‐course and extent of bone penetration before launching a clinical effectiveness trial is important. The aim of this chapter is to review the pharmacokinetics (PK) and pharmacodynamics (PD) of antibiotics in bone and present methods that support optimized evidence‐based selection of antibiotic dosage regimens.

Pharmacokinetics The time‐course and magnitude of drug concentrations in the body and particularly at the site of action determine the drug effects. Therefore, it is important to study PK, which describes the relationship between the dose of a drug and the resulting time‐course of drug concentrations at various spaces in the body  [3,4]. PK processes include drug absorption from the site of administration into the systemic circulation (except if administered directly into the bloodstream), distribution from the systemic circulation into tissues, and elimination via metabolism, renal excretion, or both. Most frequently PK is characterized based on drug concentrations measured in plasma or serum. However, in treating bone infections, adequate antibiotic concentrations must be achieved at the site of infection in bone. Numerous clinical studies have been conducted to quantify antibiotic concentrations in bone.

Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

81

82

Bone and Joint Infections

Bone is a heterogeneous tissue, where the organic bone matrix represents 30–35% of total bone mass and includes collagen fibrils (~90%), glycoproteins, proteoglycans, and extracellular fluid. Blood vessels in bone are located in Haversian and Volkmann’s canals that transverse the bone matrix. Bone cells represent only 1–2% of total bone mass and in their most mature form as osteocytes are trapped inside the bone matrix. The inorganic matrix (65–70%) consists of calcium phosphate crystals (hydroxyapatite) deposited inside the organic matrix. Due to this heterogeneous composition, most likely neither bacteria nor antibiotics distribute evenly throughout the bone tissue. The site of the pathogens in bone is not well known. Based on their size (e.g., ~1 μm for Staphylococcus aureus), bacteria are expected to distribute through the Haversian and Volkmann’s canals (~70 μm diameter) in bone, but not into the hydroxyapatite crystals. S. aureus can enter into and survive in osteoblasts, which may explain relapses. In addition, this pathogen adheres to components of the bone matrix such as collagen [5,6]. Techniques to separate the many different components of bone and measure concentrations in each are lacking. Therefore, the vast majority of published studies are based on homogenized bone samples, and the total drug concentrations in bone homogenate are reported. For the interpretation of bone penetration results, it is important to note that only free (unbound) drug is microbiologically active. However, total drug concentrations in bone homogenate, provided they are reliably determined and analyzed by population PK modeling and Monte Carlo simulations, may be more predictive of therapeutic success than serum concentrations, as the latter are further removed from events in the bone matrix.

Bone Sample Preparation and Analysis In contrast to plasma or serum, there is no specific guidance available for drug analysis in bone or other tissues. However, validated accurate and reproducible sample preparation and drug determination procedures are undoubtedly critical. When interpreting the results of published studies, it is important to consider the analytical techniques used. After bone resection, adhering blood and soft tissue is often removed from the sample. Excess blood due to intraoperative soaking can result in biased results, for example artificially high bone concentrations for a drug with low bone penetration but high blood concentrations. Samples are usually separated into cancellous bone (the inner part of the long bones) and cortical bone. Cancellous bone has a higher degree of vascularization, a higher percentage of extravascular fluid, and a lower percentage of inorganic matrix than cortical bone, which can cause differences in antibiotic penetration. For efficient extraction of the antibiotic, bone samples need to be homogenized. When bone samples are pulverized under liquid nitrogen in a cryogenic mill, this provides a very fine powder; this procedure is highly reproducible and is applicable to thermally unstable drugs (e.g., β‐lactam antibiotics) that are prone to degradation during grinding without freezing. Therefore, this method is preferable to slicing, grinding by mortar and pestle, or using mixers without cooling, as was often applied in earlier studies before more recent technology was developed. During drug extraction from the homogenized sample, sufficient and reproducible recovery and stability of the drug need to be ensured. Calibration standards and quality control samples are necessary for accurate and precise drug determination and should be prepared in drug‐free bone powder instead of plasma, serum, or buffer. An internal calibration standard should be added to each

6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone

83

s­ ample to improve the analytical accuracy and precision. Older studies frequently determined drug concentrations by microbiological assay. Newer studies have mainly employed high‐performance liquid chromatography (HPLC) or liquid chromatography–tandem mass spectrometry (LC–MS/MS), offering improved sensitivity and specificity. These chromatographic procedures have been shown to be generally superior to bioassays when analyzing bone samples [5]. Bone penetration studies should report details on the chosen methods for sample preparation and analysis, and the recovery, bias, and precision. Concentrations in bone homogenate are typically reported as mg/kg of total bone mass. Some studies report concentrations in relation to bone volume, organic bone mass, or interstitial fluid, or correct for blood content. Potential differences in reporting need to be considered when comparing results between studies.

Pharmacokinetic Sampling and Data Analysis In recent years, some studies have used microdialysis in an attempt to define a concentration time‐course of unbound antibiotic within bone of individual uninfected subjects; some such studies are considered below. However, that technique is not without limitations and it is not known whether the concentration measured represents the concentration that may occur at the site of an infection  [7]. As noted above, most studies have involved the collection of bone samples, most commonly from people undergoing orthopedic surgery. Usually, only one bone sample can be taken per patient, and a blood sample is taken at the same time. Most studies report bone penetration as the concentration ratio between bone and serum or plasma at one time point. However, due to different kinetics of drug concentrations in plasma and bone, the concentration ratios change over time until eventually an equilibrium has been reached during the terminal phase. This phenomenon (system hysteresis) hampers the interpretation of results and comparison between drugs and studies when samples are taken at different times post dosing. A better measure for the extent of bone penetration is to calculate the area under the concentration–time curve (AUC) in bone and compare it to the AUC in plasma or serum. This approach considers the full time‐course of the concentration profiles in bone and plasma (or serum). Instead of collecting the bone and blood samples at the same time point after the dose for all patients, samples should be spread out over a time period to support PK modeling. Based on such study designs, investigators have averaged the concentrations at each time point and derived the average AUCs in bone and plasma (naive averaging) [8–10]. Alternatively, one PK function has been fit to the concentration–time data from all patients (naive pooling) and the AUC integrated [11]. While these approaches remove the issue of time‐dependent concentration ratios, they only consider the average concentration–time profile and ignore the true biological variability between patients. Population PK analysis is the most powerful approach for the analysis of sparse data (e.g., one bone sample per patient) and accounts for the average rate and extent of bone penetration and interpatient variability [7,12–16]. A recent advance has been the use of a physiologically based population PK model to describe measured concentrations of ciprofloxacin in plasma and bone. This type of modeling incorporates consideration of different compartments and blood flows within bone [7]. By fitting each patient’s data in the perspective of the concentrations from the other patients, the most likely concentration– time‐course in bone and serum and the AUC can be predicted for each patient. Estimating the rate of bone penetration enables recommendations on the administration time of

84

Bone and Joint Infections

antibiotic prophylaxis before surgery. An existing population PK model can also be used to identify the optimal timing of bone and plasma samples in future bone penetration studies. Bone penetration is usually studied in joint replacement patients with uninfected bone as such patients are more easily recruited than osteomyelitis patients. The condition of the bone samples is likely more homogeneous among joint replacement patients than patients with various stages and locations of bone infections; therefore, results of different studies can be more readily compared. However, antibiotic concentrations might differ between infected and uninfected bone. Reactive hyperemia could increase the blood flow into bone, whereas pus or sequesters might limit the distribution of antibiotics into bone. To date, few studies have been performed in patients with bone infections, which do not enable a systematic comparison of penetration between infected and uninfected bone. Presence of ischemic, calcified, or arthritic tissues, bone cysts, or fat in the cancellous bone may affect antibiotic distribution. Different types of bone (e.g., hip, knee, sternum) and influences on blood circulation, for example tourniquet application or internal mammary artery harvesting, may also affect antibiotic bone concentrations.

Penetration of Antibiotics into Bone

Median bone/serum concentration ratio

Figure 6.1 presents an overview of the extent of bone penetration by antibiotic or antibiotic group. Each symbol represents the median or average bone‐to‐serum (or plasma) concentration ratio from one clinical study, and the lines indicate the median per antibiotic

3 2 1 0.5 0.3 0.2 0.1

qu in o m lon ac es cl roli in de da s m rif yci am n pi c gl lin in yc ez o o Ppe lid la p ct am p tide as enic s e i ce -inh llins ph ib al ito os rs p fo ori sf ns om fu si yci d n te ic a tra ci cy d cl in es

0.05

Figure 6.1.  Bone penetration for different antibiotic groups [5]. Each symbol represents the median or average bone‐to‐serum or bone‐to‐plasma concentration ratio from one clinical study. A concentration ratio of 11.6 for levofloxacin from one study is off the scale [14]. The lines represent the group medians. A comprehensive reference list for papers up to and including the year 2012 can be found in reference [5]; more recent papers are listed at the end of this chapter.

6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone

85

or antibiotic group. In total, 140 studies (until July 2020) were included. Most concentration ratios were reported directly in the published studies; sometimes they were calculated from the reported concentrations or read from plots. Tables 6.1 and 6.2 list the range of average concentration ratios for various antibiotics; for some studies dispersion around the average is provided  [7,8,10–12,14–38]. Unless otherwise indicated, concentration ratios are based on total concentrations in bone homogenate and serum (plasma); microdialysis sampling of bone was used in a few studies. Some studies are discussed in more detail in the text. Table 6.1.  Bone penetration of selected fluoroquinolones and macrolides. Range of Range of average bone/ time since serum concentration last dose ratios

Bone or surgery type

Bio‐analytical method

uninfected

0.5–13 h

0.27–1.2

Hip, knee, skull,

HPLC [10,17,18]

uninfected

0.5–20 h

Increased over time; organic bone matrix to plasma partition coefficients were 3.39 for cortical and 5.11 for cancellous bone; based on physiologically based population PK modeling

Hip replacement, fracture of femoral neck, humerus, femur, tibia, pelvic column

HPLC [7]

osteomyelitis

2–4.5 h

0.42

debridement surgery

HPLC [17]

uninfected

0.7–2 h

0.36–1.0

Hip, other

HPLC [19–21]

uninfected

1–3 h

3.8 ± 2.1 at 1 h, 11.6 ± 6.4 at 2 h, 20.9 ± 11.9 at 3 h; based on simulations from population PK modeling

Hip or knee replacement; cortical bone analyzed

HPLC [14]

0.5–12 h

0.09–1.04

Hip, nasal bone, mastoid process

HPLC [22–24]

1.5–5 h

0.33–1.05

Hip, knee, sternum, manubrium

HPLC [21,25–27]

0.5–6.5 days

2.5–6.3

Alveolar bone

Bioassay [28,29]

3.3–24 h

1.5–2.6

Ethmoid bone

HPLC [8]

Antibiotic and bone condition Ciprofloxacin

Levofloxacin

Ofloxacin uninfected Moxifloxacin uninfected Azithromycin uninfected Telithromycin uninfected

HPLC, high‐performance liquid chromatography.

Table 6.2.  Bone penetration of selected beta‐lactams. Antibiotic and bone condition

Range of time since last dose

Range of average bone/serum concentration ratios

Bone or surgery type

Bio‐analytical method

0.5–6 h

0.03–0.31

Hip, jaw

bioassay [11,30,31], LC‐MS/MS [32]

0.5–6 h

0.01–0.14

Hip

bioassay [11,30], LC‐MS/MS [32]

1–2 h

0.46–0.76a

Hip

HPLC [33]

Uninfected

2h

0.54

Cardiac surgery

Bioassay [34]

Ischemic bone

1–2 h

0.04–0.08

foot

HPLC [35,36]

2.5–24 h (microdialysis)

0.74 ± 0.36 L sternum, 0.99 ± 0.59 R sternum; based on AUCs

Sternal cancellous bone

HPLC [37]

1–5 h

0.08 ± 0.04 at 1 h, 1.12 ± 1.29 at 5 h; based on simulations from population PK modeling

Hip or knee replacement; cancellous bone

HPLC [12]

Amoxicillin Uninfected Clavulanic acid Uninfected Cefepime Uninfected Ceftazidime

Cefazolin Uninfected Ceftriaxone Uninfected

Cefuroxime Uninfected

0.25–8 h (microdialysis)

short‐term infusion: cancellous bone 1.03, cortical bone 0.35 continuous infusion: cancellous bone 1.15, cortical bone 0.65 based on unbound AUCs after population PK modeling

Knee replacement; microdialysis cancellous and cortical tibia bone

HPLC [15]

2–28 h

0.025; from population PK modeling

Joint replacement or lower limb amputation

LC‐MS [16]

3–71 min prior to skin incision

hip replacement: 0.078 femoral head; 0.071 femoral neck   knee replacement: 0.056 femur; 0.055 tibia

Hip or knee replacement

HPLC [38]

Ertapenem Uninfected Flucloxacillin Uninfected

HPLC, high‐performance liquid chromatography; LC‐MS and LC–MS/MS, liquid chromatography–mass spectrometry and liquid chromatography– tandem mass spectrometry. a Assuming a bone density of 1 kg/L.

88

Bone and Joint Infections

The data summarized in Figure 6.1 reveal some systematic differences across antibiotic groups, which may be due to different physicochemical and binding characteristics. There was an approximate five‐fold difference in the median bone/serum concentration ratio of the antibiotic class with the lowest ratio (penicillins) and the class with the highest ratio (macrolides). Median bone‐to‐serum concentration ratios were 0.50 for quinolones and 0.48 for linezolid. Despite large differences in chemical structure, clindamycin, rifampicin, glycopeptides, fosfomycin, and fusidic acid had comparable median concentration ratios of 0.23–0.35. Penicillins, cephalosporins, and β‐lactamase inhibitors showed median concentration ratios of 0.16, 0.21, and 0.22. Figure 6.1 also demonstrates large variability within each antibiotic group. This is most noticeable for the macrolides for which median ratios spanned a very wide range. Part of this variability may relate to the differing physicochemical characteristics across the macrolides as the ratio for azithromycin was reported to be substantially higher than that for spiramycin [5]. Also contributing to the variable ratios reported may be differences in analytical procedures (e.g., some studies included in Figure 6.1 used bioassays while others used liquid chromatography methods) and, as discussed, collection of samples at different times after drug administration.

Fluoroquinolones Fluoroquinolones are frequently used in bone infections, and show one of the highest median extents of bone penetration of all antibiotic groups with bone‐to‐serum concentration ratios mostly between 0.3 and 1.2 (Figure 6.1 and Table 6.1). The high penetration may be partly due to binding of quinolones to the calcium in bone. As only free antibiotic is considered microbiologically active, the quinolone concentrations available for antimicrobial action are likely lower than the total bone concentrations. The concentration ratios of most quinolones tend to increase with time since the last dose, as recently reported for ciprofloxacin, indicating slow redistribution from bone back into the bloodstream [7]. Quinolones generally penetrate well into cells. This could be advantageous for treatment of S. aureus osteomyelitis, since S. aureus was shown to penetrate into and survive in osteoblasts in vitro [6]. Multiple studies in different patient groups have examined the bone penetration of ciprofloxacin (Table 6.1). In the most recently reported study, which was conducted in uninfected patients undergoing orthopedic surgery (n = 39), the average observed cortical bone‐to‐plasma concentration ratio was 0.67 at 0.5 to 2 h and 5.1 at 13 to 20 h; for cancellous bone the respective average ratios were 0.77 and 4.4 [7]. The novel physiologically based population PK modelling in that study estimated the partition coefficient of the organic bone matrix was 3.39 for cortical and 5.11 for cancellous bone. Fong et al. [17] compared ciprofloxacin concentrations in cortical bone from patients without (n = 18, hip or knee replacement or osteotomy) and with (n = 10) osteomyelitis. Concentrations in infected bone were 30–100% higher than in uninfected bone. As serum concentrations were also higher in osteomyelitis patients, the average bone‐to‐serum concentration ratios were approximately 0.4 in both patient groups. Bone penetration of levofloxacin was evaluated in four studies (Table 6.1). In patients undergoing bone surgery (n = 9) or decubitus ulcer debridement (n = 12), the bone‐to‐ serum concentration ratios were 0.36 for cortical (n = 6) and 0.85 ± 0.40 for cancellous (n = 14) bone [19]. In 12 hip replacement patients, ratios of 1.0 ± 0.4 for cortical and

6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone

89

0.5 ± 0.1 for cancellous bone were reported [20]. Concentration ratios of 0.42 ± 0.04 in cortical and 0.54 ± 0.05 in cancellous bone were found in eight hip replacement patients [21]. In a more recent study, substantially higher cortical bone‐to‐plasma concentration ratios were reported, the explanation for which remains to be determined [14]. Moxifloxacin was studied in four studies, which showed consistently high penetration considering the range of different bone types (Table 6.1) and methods for sample homogenization (hand mincing, sonication, cryogenic mill). In all moxifloxacin studies, the penetration into cancellous and cortical bone was similar. Utilizing a cryogenic mill and population PK analysis, the bone‐to‐serum AUC ratios in 24 hip replacement patients were 0.80 (10th–90th percentile for between‐patient variability: 0.51–1.26) for cortical and 0.78 (0.42–1.44) for cancellous bone [25].

Macrolides and Telithromycin Studies on macrolides, some of which were conducted decades ago using bioassays, demonstrate the largest range of penetration of all antibiotic groups (Figure 6.1). In 2 more recent studies  [28,29], patients received 500 mg azithromycin once daily for 3 days before periodontal surgery. In both studies, the average concentration ratios increased slightly from 12 h to 2.5 days, when they reached greater than 6, and then slowly decreased to approximately 2.5 at 6.5 days. The rate of azithromycin penetration into bone is not known, as the first samples were taken at 12 h. Telithromycin penetration into ethmoid bone was studied in 29 patients [8]. The concentration ratio increased between 3 and 24 h, indicating relatively slow equilibration that may favor administration at least approximately 12 h ahead of surgery. Using naive averaging, the average bone‐to‐serum AUC ratio was 1.6, being one of the highest extents of penetration of all studied antibiotics.

Clindamycin Clindamycin is often referred to as possessing high bone penetration. The median bone‐to‐serum concentration ratio of 0.35 from four clindamycin studies, however, appears to be lower than for fluoroquinolones (median 0.50) and linezolid (median 0.48) (Figure 6.1). As most clindamycin studies were performed in the 1970s, that is before the introduction of fluoroquinolones, linezolid, and azithromycin, the bone penetration of clindamycin was higher than that of other available antibiotics at that time. All available clindamycin studies used bioassays, which have the potential to be confounded by active metabolites of clindamycin. Results from multiple studies suggest an extent of bone penetration of clindamycin of 0.21–0.45, similar to or slightly higher than cephalosporins [5].

Rifampicin A wide range of bone‐to‐serum concentration ratios (0.08–0.56 at 2–14 h after the dose) was found for rifampicin in four studies in uninfected bone from the 1970s/1980s. One of the studies also investigated infected bone, and concentration ratios were similar to those for uninfected bone (0.57 versus 0.46). All studies utilized bioassays and had high interpatient variability [5].

90

Bone and Joint Infections

Tetracyclines and Tigecycline Few studies are available for tetracyclines, and results vary despite the high binding affinity of tetracyclines to calcium [5]. For tigecycline, initial analysis of bone samples from 25 uninfected surgical patients suggested relatively low concentrations [9]. Re‐analysis of the same samples by a new LC–MS/MS assay, including a stabilizing agent, resulted in bone concentrations that were on an average 9.5‐fold higher as compared to the previous method  [39]. This demonstrates the importance of validating all aspects of analytical methods. Extensive penetration of tigecycline into bone has been reported in a more recent study in 33 uninfected surgical patients. Using a validated LC‐MS/MS assay the authors reported a bone‐to‐serum concentration ratio of 4.77 based on the ratio of AUCs in the matrices [40].

Cephalosporins Numerous studies have been performed with cephalosporins. For cefuroxime, the average bone‐to‐serum concentration ratio was 0.32 (range 0.09–0.55, 10 min to 6.5 h post dose) in five studies that reported concentrations in serum and uninfected bone and in which the majority of samples were above the detection limit [5]. In a recently reported study, bone penetration of cefuroxime was studied in uninfected surgical patients undergoing knee replacement [15]. In 9 patients who received the drug by short‐term infusion and in 9 patients administered a continuous infusion, the respective bone‐to‐plasma concentration ratios based on unbound AUCs after population PK modeling were 1.03 for cancellous bone and 0.35 for cortical bone; the corresponding ratios for continuous infusion were 1.15 and 0.65 (Table 6.2). Ceftriaxone and cefamandole were evaluated within the same study in hip replacement patients  [41]. At 10–30 min after the dose, average (95% confidence interval) bone‐to‐ serum concentration ratios were 0.156 (0.123–0.190) for ceftriaxone and 0.184 (0.156– 0.212) for cefamandole. The bone‐to‐plasma concentration ratios based on total drug were similar despite a six‐fold difference in the non‐protein‐bound fractions in plasma (0.05 for ceftriaxone, 0.30 for cefamandole), although only unbound drug is believed to distribute between plasma and tissues. This issue is discussed in more detail in our previous review [5]. At 8 h after the dose, ceftriaxone bone‐to‐serum concentration ratios were 0.142 (0.073–0.210), very similar to those at 10–30 min, suggesting a fast equilibration between serum and bone [41]. However, as noted in Table 6.2, a more recent study suggested a slower equilibration [12]. In 11 patients undergoing debridement for septic non‐ union of the tibia, average bone‐to‐plasma AUC ratios of ceftriaxone were 0.093 in cortical and 0.241 in cancellous bone [42]. Average bone‐to‐serum concentration ratios for cefamandole in hip replacement patients were 0.227–0.249 at 10–30 min after the dose, and the overall range of concentration ratios reported was 0.12–2.3 at 10 min to 4 h  [5]. For cefazolin, median bone‐to‐ serum concentration ratios in eight infected patients were 0.25 (range 0.06–0.41) during continuous infusion, with concentrations determined by bioassay [43]. In a more recent study involving bone microdialysis and plasma sampling over 24 h, and quantification of cefazolin by HPLC, average bone‐to‐plasma concentration ratios based on AUCs of 0.74 ± 0.36 and 0.99 ± 0.59 were observed for left and right sternum of patients undergoing coronary artery bypass grafting (CABG) [37]. Additional results for several β‐lactams are presented in Table 6.2.

6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone

91

Overall cephalosporins achieved concentration ratios of 0.1–1. Penetration was higher into cancellous bone than into cortical bone in all studies that analyzed both, potentially due to the higher proportion of extracellular fluid in cancellous bone  [5]. β‐Lactams, including cephalosporins, are assumed to distribute mainly into extracellular fluid and were found to exhibit limited binding to the inorganic bone matrix.

Penicillins, Carbapenems, and β‐Lactamase Inhibitors Two studies by different groups from 1994 and 2001 evaluated piperacillin/tazobactam penetration into uninfected hip bone in 12 patients each, used the same sample preparation methods and analysis by HPLC, and found consistent results. Bone‐to‐plasma concentration ratios were 0.2–0.3 for piperacillin and tazobactam in cortical and cancellous bone at 1–1.5 h after the dose [44,45]. More recently, penetration of piperacillin/tazobactam into uninfected jaw (n = 7) and hip (n = 2) bone was studied [46]. Sample preparation was similar to the previous studies and concentrations were analyzed by LC‐MS/MS. At an average of 3 h (range 1–7 h) after the start of the infusion, bone‐to‐plasma concentration ratios were 0.15 for piperacillin and 0.13 for tazobactam. These results were slightly lower and more variable than those from the previous studies, potentially due to different bone types and the wider range of sampling times. A wide range of average amoxicillin bone‐to‐serum concentration ratios was reported in studies utilizing bioassays (Table 6.2). In a study in 20 hip replacement patients analyzed by LC‐MS/MS and population PK analysis, the bone‐to‐serum AUC ratios were 0.20 (10th–90th percentile for between‐ patient variability 0.16–0.25) for cortical and 0.18 (0.11–0.29) for cancellous bone [32]. In the same study, the bone‐to‐serum AUC ratios of clavulanic acid were 0.15 (0.11–0.21) for cortical and 0.10 (0.051–0.21) for cancellous bone. A recent report has documented bone‐to‐serum concentration ratios of flucloxacillin, based on HPLC analysis, of 0.07– 0.08 and 0.05–0.06 for hip and knee bone, respectively, in the first 1.5 h after drug administration (Table 6.2) [38]. Carbapenems have been the focus in only a small number of studies, and interpretation has been confounded by methodological problems  [5]. However, a recent study on ertapenem involving sampling over 2–28 h, analysis by LC‐ MS and population PK modeling reported a bone‐to‐plasma concentration ratio of 0.025 in 10 patients [16].

Linezolid Linezolid is relatively stable, as opposed to many β‐lactams, and the available studies were performed utilizing HPLC. Average (95% confidence interval) bone‐to‐serum concentration ratios were 0.51 (0.43‐0.75) in 12 hip replacement patients at 30–50 min after the start of the infusion [47]. Similar penetration (0.40 ± 0.24) was also found at 1.5 h in 12 elderly patients during knee replacement [48]. At the same dose as the two joint replacement studies [47, 48], and 0.5–1.5 h after the dose, lower linezolid concentrations were found in 11 patients with implant‐associated infections. The average bone‐to‐plasma concentration ratio was approximately 0.23 [49]. However, in a recent study involving analysis of linezolid by LC‐MS/MS in 9 patients with spinal tuberculosis, the average (range) bone‐to‐plasma concentration ratio 24 h after drug administration was 0.48 (0.30– 0.67) [50]. Two studies have analyzed linezolid in bone by microdialysis [37,51]. The ratio of unbound AUCs (fAUC) in cancellous bone/plasma over 12 h was 1.09 ± 0.11 in 3 diabetic patients with severe foot infections [51], while over 24 h in 9 uninfected patients

92

Bone and Joint Infections

undergoing CABG the ratio for sternal bone was 0.82 ± 0.28 for the left sternum and 1.02 ± 0.47 for the right sternum [37]. The finding of ratios close to unity in both studies indicates similar exposure to microbiologically active linezolid in interstitial bone fluid and plasma  [37,51]. The higher AUC ratio from microdialysis, compared to concentration ratios based on bone homogenate, is in keeping with a low propensity of linezolid to form chelate complexes with the inorganic bone matrix.

Daptomycin Bone penetration of daptomycin was evaluated in diabetic foot infections  [52]. Serial microdialysis samples at steady state were collected from 0 to 8 h after the dose in 5 patients and from 8 to 16 h in another 4 patients. The average ratio of the fAUC (0–16 h) in interstitial fluid of metatarsal bone/plasma was 1.08. This ratio of unbound bone‐to‐ unbound plasma concentrations suggests high penetration of daptomycin into interstitial fluid, that is, the most likely site of infection, and was achieved for a drug with high plasma protein binding (~90%) and a high molecular weight. Indeed, the high plasma protein binding is the likely reason why a study that measured the ratio of total AUCs reported a substantially lower median concentration ratio of 0.095 for thigh bone and 0.082 for shin bone [53]. This highlights the importance of considering protein binding when interpreting the possible clinical consequences of bone penetration results.

Fosfomycin Bone homogenate‐to‐serum concentration ratios of 0.13–0.45 were reported in three studies from 1980 to 1983 utilizing bioassays (Figure  6.1). Fosfomycin binds to hydroxyapatite in bone, suggesting that not all fosfomycin in bone homogenate is microbiologically active. However, a microdialysis study found an average fAUC ratio of 0.43 ± 0.04 in nine osteomyelitis patients with diabetic foot infection, which is higher than or similar to the reported bone homogenate‐to‐plasma concentration ratios [54]. Considering the very limited or no binding of fosfomycin to plasma proteins, this would suggest that the average concentration bound to various components of bone tissue is lower than or similar to the interstitial fluid concentrations. However, comparison among studies is hampered by differences in study designs and methodologies.

Glycopeptides and Lipoglycopeptides A wide range of average concentration ratios, mostly between 0.1 and 0.6, has been reported for glycopeptides in hip, knee, or sternal bone (Figure 6.1). A recent study used microdialysis sampling of cancellous and cortical bone to examine bone penetration of vancomycin in 10 patients undergoing knee replacement surgery [55]. The median (95% confidence interval) fAUC bone/plasma ratio based on samples collected up to 8 h was 0.45 (0.29‐0.62) for cancellous bone and 0.17 (0.11‐0.24) for cortical bone. The bone penetration of dalbavancin was studied in 30 patients undergoing knee or hip replacement surgery [13]. Dalbavancin was quantified in cortical bone and plasma using LC‐MS/MS, and the resulting data subjected to population PK modeling. The bone‐to‐plasma concentration ratio was 0.131. When interpreting this value, it is important to recognize that dalbavancin is approximately 90–95% bound in plasma.

93

6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone

Pharmacodynamics and Monte Carlo Simulations PD describes the relationship between drug concentrations in plasma or at the target (i.e., infection) site and the time‐course of drug effect(s). For β‐lactams, the time during which the unbound antibiotic concentration remains above the minimum inhibitory concentration (fT> MIC) of the pathogen has been shown to be predictive of the extent of antibiotic effect. For other antibiotics, such as fluoroquinolones, the fAUC/MIC best correlates with effect. For bone penetration studies that used the same sampling time for all samples, it is not possible to perform a PD analysis because comparing the bone concentration at a specific time to the MIC provides limited information. Irrespective of the type of data analysis used, adequate reporting of the methods and assumptions is important. Naive pooling or averaging approaches, as described earlier, can calculate the average AUC/MIC and time above MIC. This indicates whether an “average” patient would attain the PK/PD target. However, these naive methods have the disadvantage that they do not consider the true variability between patients, which tends to be large for bone penetration. Population modeling accounts for both the average penetration and its variability between subjects (Figure  6.2). Once a population PK model for plasma and bone has been developed, it can be employed in Monte Carlo simulations to predict the expected concentration‐time profiles for other than the studied dosage regimens. This includes Serum

5 4 3 2 1 0

0

1

2

3

4

Cortical bone

6 Concentration (mg/kg)

Concentration (mg/L)

6

5

6

7

5 4 3 2 1 0

0

1

2

3

Time (h) Concentration (mg/kg)

5

6

7

Cancellous bone

7 6

P25% percentile

5

P5% percentile

4 3

P50% percentile P75% percentile

2

P95% percentile

1

Raw data

0

4

Time (h)

0

1

2

3

4

5

6

7

Time (h)

Figure 6.2.  Visual predictive checks for the moxifloxacin population model presenting both the median concentration–time profiles and their between‐subject variability in serum, cortical bone, and cancellous bone. The diamonds represent the individual observed data from 24 hip replacement patients. The solid lines represent the model‐predicted median profiles, and the dashed lines are the model‐predicted 5, 25, 75, and 95% percentiles. (From Ref. [25], © American Society for Microbiology).

94

Bone and Joint Infections

predicting the variability in concentration‐time profiles between patients. In this way, the probability of achieving a PK/PD target can be predicted and recommendations made on how to dose an antibiotic to maximize the probability of successful therapeutic outcome. The PK/PD target values for plasma and bone concentrations to successfully treat bone infections are most often unknown, and the target values for other types of infections can likely not be used. For moxifloxacin, no published clinical studies in osteomyelitis were available. To address the lack of known PK/PD target values for bone, a reverse engineering approach was applied for moxifloxacin to identify the most likely PK/PD target required to achieve clinical and microbiological cure of osteomyelitis  [25]. This approach combined effectiveness data from clinical studies with ciprofloxacin in osteomyelitis, the expected plasma AUCs from these studies, the AUCbone‐to‐AUCplasma ratio for ciprofloxacin [10], fraction free (unbound) in plasma, and bacterial susceptibility data from the time of the clinical studies. Reverse engineering suggested a fAUC/MIC of 40 in serum and an AUC/MIC of 33 in bone as the most likely PK/PD targets for successful clinical and microbiological outcome. No assumptions were made regarding the numerical value of the free fraction of moxifloxacin in bone. It was assumed that binding and distribution within the bone tissue is similar for moxifloxacin and ciprofloxacin, two quinolones with the same essential chemical structure that are expected to be responsible for binding characteristics. The population PK model for moxifloxacin in serum and bone was utilized to predict likely probabilities of target attainment. An ≥90% probability of successful clinical and microbiological outcome was predicted for 400 mg moxifloxacin once daily up to an MIC of 0.375 mg/L (mg/kg) in serum and cancellous bone and 0.5 mg/L in cortical bone (Figure 6.3). Compared to, for example, an MIC90 of 0.125 mg/L for S. aureus, these are favorable results and suggest clinical trials are warranted. The antibiotic susceptibility of the local hospital should be considered when published probabilities of target attainment are used to decide about antibiotic therapy in patients. Similar analyses to determine the probability of target attainment using clinically safe dosing regimens have been conducted for amoxicillin/clavulanic acid [32] and ciprofloxacin [7]. In the case of ciprofloxacin, the analysis utilized the first reported physiologically based population PK model to employ measured concentrations of the drug in plasma and cortical and cancellous bone. The population PK and Monte Carlo simulation approach described for bone is applicable to other matrices, for example synovial fluid. For antibiotics for which clinical effectiveness trials are not available, combining population PK, the reverse engineering approach utilizing effectiveness data from literature as described earlier, and Monte Carlo simulations appears to be the best available approach currently to derive PK/PD targets for successful treatment of bone infections and suggest dosage regimens to be studied in clinical effectiveness trials. While sufficient bone penetration is an important factor, bone concentrations alone provide limited information to draw conclusions on the effectiveness of an antibiotic. Therefore, clinical recommendations should not be made exclusively based on bone penetration studies. An antibiotic also needs to have adequate antibacterial activity against the infecting pathogen. Well‐ controlled PK/PD studies in osteomyelitis patients would be required to further quantitatively elucidate the PK/PD relationship between antibiotic bone concentrations and clinical outcomes. However, such studies are currently scarce.

95

6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone

Probability of target attainment

Serum 100%

PK/PD targets for clinical and microbiological cure based on reverse 60% engineering [25]

80% 60% 40%

Serum: fAUC/MIC ≥ 40 Bone: AUC/MIC ≥ 33

20% 0% 0.125

0.25

0.5

1

2

4

8

MIC (mg/L)

Cancellous bone Probability of target attainment

Probability of target attainment

Cortical bone 100% 80% 60% 40% 20% 0% 0.125

0.25

0.5

1 MIC (μg/g)

2

4

8

100% 80% 60% 40% 20% 0% 0.125

0.25

0.5

1

2

4

8

MIC (μg/g)

Figure 6.3.  Probabilities of target attainment to achieve successful clinical and microbiological outcome. The PK/PD targets for serum, cortical bone, and cancellous bone are based on a reverse engineering approach [25].

Conclusions Trends in the extent of bone penetration among different groups of antibiotics have been found from a review of greater than 140 literature studies, such as typical average penetration of 0.3–1.2 for fluoroquinolones, 0.2–0.5 for linezolid, 0.1–0.3 for penicillins, and 0.1–0.5 for cephalosporins. These differences are most likely due to different physicochemical characteristics of the antibiotic groups. For several antibiotics, measured concentrations are slightly higher in cancellous bone than those in cortical bone. Developing approaches that provide greater insights into distribution and binding of antibiotics within bone is expected to be clinically useful. High variability between studies for a particular antibiotic group is likely partly due to a lack of standardization of bioanalytical methods and study designs. The variability between patients within a study needs to be considered, and this can be achieved by population PK modeling. More data are needed to characterize the effect of infected versus uninfected bone, the presence of an implant, and the type of bone (e.g., hip, knee, sternum) on the PK within bone. Future clinical studies should focus on validated bioanalytical methods, as well as study designs, and apply PK/PD analyses that incorporate the time‐course of bone concentrations to contribute to evidence‐based care for patients with bone infections.

96

Bone and Joint Infections

Key Points ●

●

●

There are differences in the extent of bone penetration among various antibiotic groups. These trends are likely related to different physicochemical and pharmacokinetic characteristics. The variability within antibiotic groups and between different studies for the same agent is high. Therefore, utilizing standardized, validated methods for sample preparation and analysis, as well as calculating the bone‐to‐serum AUC ratios instead of single time‐point concentration ratios, are essential. Based on the currently available data, combining population PK modeling, effectiveness data from the literature, and Monte Carlo simulations appears to be the most promising approach to elucidate the extent and time‐course of bone penetration and its relationship with likely clinical outcomes. Well‐controlled PK/PD studies in osteomyelitis patients are required to directly identify PK/PD targets.

References 1. Kurtz SM, Lau EC, Son MS, et al. Are we winning or losing the battle with periprosthetic joint infection: Trends in periprosthetic joint infection and mortality risk for the Medicare population. J Arthroplasty. 2018;33(10):3238–3245. 2. Sendi P, Zimmerli W. Antimicrobial treatment concepts for orthopaedic device‐related infection. Clin Microbiol Infect. 2012;18(12):1176–1184. 3. Tozer TN, Rowland M. Essentials of Pharmacokinetics and Pharmacodynamics. 2nd ed., Philadelphia, Wolters Kluwer, 2015. 4. Drusano GL. Pharmacokinetics and pharmacodynamics of antimicrobials. Clin Infect Dis. 2007;45 Suppl 1:S89–95. 5. Landersdorfer CB, Bulitta JB, Kinzig M, et al. Penetration of antibacterials into bone: pharmacokinetic, pharmacodynamic and bioanalytical considerations. Clin Pharmacokinet. 2009;48(2):89–124. 6. Jevon M, Guo C, Ma B, et  al. Mechanisms of internalization of Staphylococcus aureus by cultured human osteoblasts. Infect Immun. 1999;67(5):2677–2681. 7. Landersdorfer CB, Kinzig M, Höhl R, et  al. Physiologically based population pharmacokinetic modeling approach for ciprofloxacin in bone of patients undergoing orthopedic surgery. ACS Pharmacol Transl Sci. 2020;3(3):444–454. 8. Kuehnel TS, Schurr C, Lotter K, et al. Penetration of telithromycin into the nasal mucosa and ethmoid bone of patients undergoing rhinosurgery for chronic sinusitis. J Antimicrob Chemother. 2005;55(4):591–594. 9. Rodvold KA, Gotfried MH, Cwik M, et al. Serum, tissue and body fluid concentrations of tigecycline after a single 100 mg dose. J Antimicrob Chemother. 2006;58(6):1221–1229. 10. Massias L, Buffe P, Cohen B, et  al. Study of the distribution of oral ciprofloxacin into the mucosa of the middle ear and the cortical bone of the mastoid process. Chemotherapy. 1994;40 Suppl 1:3–7. 11. Weismeier K, Adam D, Heilmann HD, et  al. Penetration of amoxycillin/clavulanate into human bone. J Antimicrob Chemother. 1989;24 Suppl B:93–100. 12. Gergs U, Clauss T, Ihlefeld D, et al. Pharmacokinetics of ceftriaxone in plasma and bone of patients undergoing hip or knee surgery. J Pharm Pharmacol. 2014;66(11):1552–1558. 13. Dunne MW, Puttagunta S, Sprenger CR, et al. Extended‐duration dosing and distribution of dalbavancin into bone and articular tissue. Antimicrob Agents Chemother. 2015;59(4): 1849–1855.

6  Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone

97

14. Gergs U, Ihlefeld D, Clauss T, et al. Population pharmacokinetics of levofloxacin in plasma and bone of patients undergoing hip or knee surgery. Clin Pharmacol Drug Dev. 2018;7(7): 692–698. 15. Tøttrup M, Søballe K, Bibby BM, et al. Bone, subcutaneous tissue and plasma pharmacokinetics of cefuroxime in total knee replacement patients – a randomized controlled trial comparing continuous and short‐term infusion. APMIS. 2019;127(12):779–788. 16. Chambers J, Page‐Sharp M, Salman S, et al. Ertapenem for osteoarticular infections in obese patients: a pharmacokinetic study of plasma and bone concentrations. Eur J Clin Pharmacol. 2019;75(4):511–517. 17. Fong IW, Ledbetter WH, Vandenbroucke AC, et al. Ciprofloxacin concentrations in bone and muscle after oral dosing. Antimicrob Agents Chemother. 1986;29(3):405–408. 18. Leone M, Sampol‐Manos E, Santelli D, et al. Brain tissue penetration of ciprofloxacin following a single intravenous dose. J Antimicrob Chemother. 2002;50(4):607–609. 19. von Baum H, Böttcher S, Abel R, et al. Tissue and serum concentrations of levofloxacin in orthopaedic patients. Int J Antimicrob Agents. 2001;18(4):335–340. 20. Rimmelé T, Boselli E, Breilh D, et al. Diffusion of levofloxacin into bone and synovial tissues. J Antimicrob Chemother. 2004;53(3):533–535. 21. Metallidis S, Topsis D, Nikolaidis J, et al. Penetration of moxifloxacin and levofloxacin into cancellous and cortical bone in patients undergoing total hip arthroplasty. J Chemother. 2007;19(6):682–687. 22. Meissner A, Borner K, Koeppe P. Concentrations of ofloxacin in human bone and in cartilage. J Antimicrob Chemother. 1990;26 Suppl D:69–74. 23. Tolsdorff P. Penetration of ofloxacin into nasal tissues. Infection. 1993;21(1):66–70. 24. Tolsdorff P. Tissue and serum concentrations of ofloxacin in the ear region following a single daily oral dose of 400 mg. Infection. 1993;21(1):63–65. 25. Landersdorfer CB, Kinzig M, Hennig FF, et al. Penetration of moxifloxacin into bone evaluated by Monte Carlo simulation. Antimicrob Agents Chemother. 2009;53(5):2074–2081. 26. Malincarne L, Ghebregzabher M, Moretti MV, et al. Penetration of moxifloxacin into bone in patients undergoing total knee arthroplasty. J Antimicrob Chemother. 2006;57(5): 950–954. 27. Metallidis S, Charokopos N, Nikolaidis J, et al. Penetration of moxifloxacin into sternal bone of patients undergoing routine cardiopulmonary bypass surgery. Int J Antimicrob Agents. 2006;28(5):428–432. 28. Malizia T, Tejada MR, Ghelardi E, et  al. Periodontal tissue disposition of azithromycin. J Periodontol. 1997;68(12):1206–1209. 29. Malizia T, Batoni G, Ghelardi E, et al. Interaction between piroxicam and azithromycin during distribution to human periodontal tissues. J Periodontol. 2001;72(9):1151–1156. 30. Grimer RJ, Karpinski MR, Andrews JM, et al. Penetration of amoxycillin and clavulanic acid into bone. Chemotherapy. 1986;32(3):185–191. 31. Akimoto Y, Kaneko K, Tamura T. Amoxicillin concentrations in serum, jaw cyst, and jawbone following a single oral administration. J Oral Maxillofac Surg. 1982;40(5):287–293. 32. Landersdorfer CB, Kinzig M, Bulitta JB, et al. Bone penetration of amoxicillin and clavulanic acid evaluated by population pharmacokinetics and Monte Carlo simulation. Antimicrob Agents Chemother. 2009;53(6):2569–2578. 33. Breilh D, Boselli E, Bel JC, et al. Diffusion of cefepime into cancellous and cortical bone tissue. J Chemother. 2003;15(2):134–138. 34. Adam D, Reichart B, Williams KJ. Penetration of ceftazidime into human tissue in patients undergoing cardiac surgery. J Antimicrob Chemother. 1983;12 Suppl A:269–273. 35. Raymakers JT, Schaper NC, van der Heyden JJ, et  al. Penetration of ceftazidime into bone from severely ischaemic limbs. J Antimicrob Chemother. 1998;42(4):543–545. 36. Raymakers JT, Houben AJ, van der Heyden JJ, et al. The effect of diabetes and severe ischaemia on the penetration of ceftazidime into tissues of the limb. Diabet Med. 2001;18(3): 229–234.

98

Bone and Joint Infections

37. Andreas M, Zeitlinger M, Wisser W, et  al. Cefazolin and linezolid penetration into sternal cancellous bone during coronary artery bypass grafting. Eur J Cardiothorac Surg. 2015;48(5): 758–764. 38. Torkington MS, Davison MJ, Wheelwright EF, et al. Bone penetration of intravenous flucloxacillin and gentamicin as antibiotic prophylaxis during total hip and knee arthroplasty. Bone Joint J. 2017;99‐b(3):358–364. 39. Ji AJ, Saunders JP, Amorusi P, et al. A sensitive human bone assay for quantitation of tigecycline using LC/MS/MS. J Pharm Biomed Anal. 2008;48(3):866–875. 40. Bhattacharya I, Gotfried MH, Ji AJ, et al. Reassessment of tigecycline bone concentrations in volunteers undergoing elective orthopedic procedures. J Clin Pharmacol. 2014;54(1):70–74. 41. Lovering AM, Walsh TR, Bannister GC, et al. The penetration of ceftriaxone and cefamandole into bone, fat and haematoma and relevance of serum protein binding to their penetration into bone. J Antimicrob Chemother. 2001;47(4):483–486. 42. Garazzino S, Aprato A, Baietto L, et al. Ceftriaxone bone penetration in patients with septic non‐union of the tibia. Int J Infect Dis. 2011;15(6):e415–421. 43. Zeller V, Durand F, Kitzis MD, et al. Continuous cefazolin infusion to treat bone and joint infections: clinical efficacy, feasibility, safety, and serum and bone concentrations. Antimicrob Agents Chemother. 2009;53(3):883–887. 44. Incavo SJ, Ronchetti PJ, Choi JH, et al. Penetration of piperacillin‐tazobactam into cancellous and cortical bone tissues. Antimicrob Agents Chemother. 1994;38(4):905–907. 45. Boselli E, Breilh D, Biot L, et al. Penetration of piperacillin/tazobactam into cancellous and cortical bone tissue. Curr Ther Res Clin Exp. 2001;62(7):538–545. 46. Al‐Nawas B, Kinzig‐Schippers M, Soergel F, et al. Concentrations of piperacillin‐tazobactam in human jaw and hip bone. J Craniomaxillofac Surg. 2008;36(8):468–472. 47. Lovering AM, Zhang J, Bannister GC, et al. Penetration of linezolid into bone, fat, muscle and haematoma of patients undergoing routine hip replacement. J Antimicrob Chemother. 2002;50(1):73–77. 48. Rana B, Butcher I, Grigoris P, et  al. Linezolid penetration into osteo‐articular tissues. J Antimicrob Chemother. 2002;50(5):747–750. 49. Kutscha‐Lissberg F, Hebler U, Muhr G, et al. Linezolid penetration into bone and joint tissues infected with methicillin‐resistant staphylococci. Antimicrob Agents Chemother. 2003;47(12): 3964–3966. 50. Li Y, Huang H, Dong W, et al. Penetration of linezolid into bone tissue 24 h after administration in patients with multidrug‐resistant spinal tuberculosis. PLoS One. 2019;14(10):e0223391. 51. Traunmüller F, Schintler MV, Spendel S, et al. Linezolid concentrations in infected soft tissue and bone following repetitive doses in diabetic patients with bacterial foot infections. Int J Antimicrob Agents. 2010;36(1):84–86. 52. Traunmüller F, Schintler MV, Metzler J, et  al. Soft tissue and bone penetration abilities of daptomycin in diabetic patients with bacterial foot infections. J Antimicrob Chemother. 2010;65(6):1252–1257. 53. Montange D, Berthier F, Leclerc G, et al. Penetration of daptomycin into bone and synovial fluid in joint replacement. Antimicrob Agents Chemother. 2014;58(7):3991–3996. 54. Schintler MV, Traunmüller F, Metzler J, et  al. High fosfomycin concentrations in bone and peripheral soft tissue in diabetic patients presenting with bacterial foot infection. J Antimicrob Chemother. 2009;64(3):574–578. 55. Bue M, Tøttrup M, Hanberg P, et al. Bone and subcutaneous adipose tissue pharmacokinetics of vancomycin in total knee replacement patients. Acta Orthop. 2018;89(1):95–100.

Chapter 7

Preclinical Models of Infection in Bone and Joint Surgery Caroline Constant, Lorenzo Calabro, Willem‐Jan Metsemakers, R. Geoff Richards, and T. Fintan Moriarty

Introduction Increasing placement of medical devices, coupled with rising antibiotic resistance amongst bacteria within the hospital and community environments, will ensure that bone and joint infection continues to pose a major challenge to clinicians across numerous medical specialties in the decades to come. Current treatment algorithms have benefited from extensive preclinical studies providing in vivo evidence with regard to antimicrobial selection and dosing. In the near future, further preclinical studies are expected to con‑ tribute vital efficacy data for new technologies such as antimicrobial loaded coatings and vaccines, as well as rapid, sensitive, and specific diagnostics. A robust preclinical assess‑ ment of any antimicrobial strategy, and safe and expedient implementation of any such technology, relies on a well‐designed and clinically relevant in vivo simulation using ani‑ mal models. Many animal models of bone and joint infection have been described in the literature. However, fully standardized, or universally accepted reference models are still lacking. The design variables involved in creating an animal model for bone and joint infection are multiple and inevitably require some compromise. In the field of orthopedic trauma, where fracture‐related infection (FRI) involves additional features not commonly associ‑ ated with other types of bone and joint infection (e.g. soft tissue and vascular damage, traumatic fracture), standardization and refinement of fixation methods is particularly important to improve existing models and increase their clinical translational potential. Biomechanically stable and repeatable fixation systems, which mirror clinical practice and allow reliable healing of fractures without complication, should be the starting point for clinically relevant research and development into anti‐infective strategies. The availa‑ bility of custom‐made fracture fixation systems for small laboratory animals [1] and the increasing awareness of the importance of replicating the key features of infection, such

Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

99

100

Bone and Joint Infections

as soft tissue involvement, in animal models are important developments in the field. Similarly, the advent of humanized mouse models is another development with signifi‑ cant potential to improve the value of preclinical in vivo studies. These refinements and others should enable the creation of robust, controlled, and consistent models, which allow strong scientific conclusions with a minimum of harm to animals.

Influence of Species in Preclinical Models of Bone and Joint Infection Many different animal species have been used in preclinical studies as surrogates for bone and joint infection in humans (Figure 7.1) [2]. There is currently no evidence that species selection significantly impacts on the validity of a given study. Most studies include small animals such as rats, rabbits, mice, and guinea pigs because of their lower cost and easier housing and handling [3]. The central premise in performing a preclinical in vivo trial is that the pathophysiological and therapeutic response in a chosen animal model is suffi‑ ciently similar to that in humans to allow valid extrapolation of findings. The ideal animal model for bone and joint infection research would (i) have molecular, cellular, structural, and mechanical features akin to human bone; (ii) have a size and temperament allowing low‐cost maintenance and handling; (iii) have a well‐documented genetic and immunological profile; and (iv) be sufficiently robust to endure medical and surgical interventions that reflect current clinical practice. In reality, many of these ­features are highly variable between species, and an effective study design requires aware‑ ness of these differences. The clinical reality of many bone and joint infection cases involves chronic infection, implant loosening, long‐term antibiotic administration, and multiple surgeries. It is clear that replicating such scenarios in preclinical animal models would inevitably result in a high burden for the animals involved, a burden that may not be justifiable in many preclinical studies. The extent to which a particular model is required to reflect clinical conditions should therefore be based on a considered approach dependent upon the research goals at hand along with consideration of the burden upon the animal. Historically, the success of animal models of bone and joint infection has been deter‑ mined by the degree to which radiographic, histological, and microbiological outcomes mirror those found in human disease. Host response to infection is dependent on the innate immunity of the host organism. The mobilization of adaptive immune defenses, and bone modeling and remodeling mechanisms, are all species‐dependent. Bone com‑ position and micro‐ and macrostructure are important determinants of its mechanical properties and vary between species, within species, and between anatomical loci of indi‑ vidual animals [4]. A study by Aerssens et al. [4] looked at the bone composition of cows, sheep, chickens, and rats. The mineral content of femoral cortical bone was sig‑ nificantly greater in all four species than in humans, with rats having the highest con‑ tent. The proportion of collagen content showed the inverse relationship, with rats having the lowest content. Microarchitecture of bone is related to remodeling charac‑ teristics and also differs between species. Human cortical bone, from the fetal stage onward, shows a high degree of remodeling with a secondary osteonal structure con‑ taining concentric layers of compact bone tissue that surround a central haversian canal. Nonhuman primates and dogs share a similar microstructure [5], while sheep, pigs, goats, and cows begin life with a plexiform bone structure, and develop only

7  Preclinical Models of Infection in Bone and Joint Surgery

Tibia

Radius

101

Tibia

Figure 7.1.  Examples of different animal species used in preclinical studies of bone and joint infection at the laboratories of the co‐authors. The sheep model of chronic osteomyelitis (left) has been used to determine the effectiveness of new local antibiotic carrier (antibiotic loaded hydrogel in light blue) with intramedullary locking nail; the rabbit radius defect model to compare the capacity of antibiotic‐eluting scaffolds to eradicate infection; and a rat tibial screw model to determine impact of comorbidities on osteolysis, antibiotic efficacy and interference with repair processes.

s­ econdary osteons in certain locations later in life  [6]. Rodents and rabbits have a primary lamellar bone structure, and secondary osteons are rare [5]. Another important requirement for an animal model of bone and joint infection is the capability of the test microorganism to cause an infection in the species under study. Human pathogens do not necessarily cause predictable disease in a particular animal or any disease at all. In addition, different animals may act differently to the same bacterial inoculum [1]. Staphylococcus aureus is the pathogen cultured most frequently from clini‑ cal osteomyelitis cases [7] and is also an important cause of infection in other mammals and domestic species [8]. Comparison of S. aureus strains from human and animal clini‑ cal isolates revealed that veterinary S. aureus infections are caused, to a large extent, by genetically and phenotypically distinct strains [9]. The potential implication is that patho‑ genic strains have evolved separately with different immunological selection pressures exerted by different immune defenses in different species. In addition to S. aureus, the

102

Bone and Joint Infections

most common cause of osteomyelitis in domestic animals from trauma and surgical site infection extending to the bone include S. pseudintermedius and Streptococcus spp [8]. Interspecies differences in immune response to bacterial invasion are also important to acknowledge in infection trials. With specific reference to the causative microorganisms involved, most animal species used in preclinical research are specific pathogen‐free, and may have limited exposure to bacterial infection, except in case of minor cutaneous abra‑ sions allowing the skin microbiome to invade. Therefore, most animals enrolled in a bone and joint infection study will have a very limited repertoire of circulating antibodies to the causative pathogen. This is in contrast to the human situation, where it is considered most people will encounter S. aureus and many coagulase‐negative staphylococci (CNS) throughout their lifetime, and consequently have a significant adaptive immune memory for many staphylococcal antigens. It remains to be determined what impact this has on the progression of infection. Its relevance in the clinical situation has not been greatly considered in preclinical research. Murine models represent an attractive option for preclinical investigation into the host response to infection due to the variety of genetically defined mouse strains, both wild type and mutant, the availability of humanized strains, and the great array of molecular different biology tools. For example, T helper (TH)–type responses are crucial in the immune response to infection, and may be closely controlled by careful selection of mouse strains. Several bacterial virulence factors, including bicomponent toxins, are specific for human receptors and represent a potential confounding factor in preclinical trials. For example, common methicillin‐resistant S. aureus (MRSA) strains causing community‐ associated (CA)‐MRSA infection express Panton–Valentine leukocidin (PVL)  [10]. Although PVL is epidemiologically linked to severe infections in humans [11], PVL null mutants do not reliably demonstrate reduced severity in murine models [12,13], although it can be significantly attenuated in rabbits [14]. Likewise, the ability of PVL to activate or kill neutrophils is known to vary between species [10]. The molecular basis for this discrepancy was attributed to the high human and rabbit specificity for the PVL recep‑ tors C5aR and C5L2 [15]. The frequent recurrence of S. aureus infection and its ability to manipulate immune responses in reaction to the pressures imposed by the human immune system is thought to be responsible for the human‐specific activity of several potentially important staphylococcal toxins [16]. This highlights the importance of species selection, since some animals do not necessarily correctly replicate all facets of S. aureus disease in humans. The recent development of immunodeficient mice reconstituted with a human immune system [17,18], commonly called humanized mice, provide a framework to assess the contribution of human‐specific toxins in preclinical research using a murine model of infection [13]. These engineered mice are first made immunodeficient via gene deletion or backcrossing strains with mutations in essential cellular partitions, such as macrophages, natural killer, and T and B cells  [17]. Subsequently, human cells and/or tissues are engrafted to recapitulate human immune responses  [13]. Despite the vast potential of humanized mice, there are substantial obstacles associated with the model that remain to be resolved to enhance transability  [17,19] and have not been substantially applied in bone and joint infection models to date. When evaluating bone and joint infection, the implant systems used in preclinical research are another important factor to consider. For example, in case of a fracture, the mechanical environment (i.e. fixation stability) is known to influence bone formation, revascularization, and infection susceptibility [1,20]. In this regard, the use of larger animals

7  Preclinical Models of Infection in Bone and Joint Surgery

103

such as nonhuman primates, sheep, goats, and dogs can represent an advantage since their bony geometry can accommodate human‐scale prostheses mimicking clinical ­scenarios  [2]. In small animals, fixation is often performed using intramedullary K‐wire [21–23], which may negatively impact the repair solidity compared to more stable fixations used in clinical settings. However, small‐scale fixation devices for rodents such as internal fixators analogous to a locking plate and interlocking nails have been devel‑ oped and have enabled small animal fracture models to better emulate clinical conditions. These models allow the investigator to evaluate fracture healing, and choose between rigid or flexible fixation to model primarily intramembranous or endochondral fracture healing [24].

Overview of Animal Models Experimental bone and joint infection models have been created in animals for diverse purposes, by a range of means and with varied results. Common goals include profiling infection parameters such as bacterial virulence factors or the performance of novel diag‑ nostic tools, interventions, or biomaterials. Infection is typically created by bacterial inoculation coupled with a local perturbation in bone physiology. This can be done by an implanted foreign body, experimentally induced ischemia, or administration of a scleros‑ ing agent. Study design in bone and joint infection research aims to reflect the clinical situation and fit within clinical classifications (see Chapter 16). Today, two important focus areas within the field of bone and joint infection research are fracture‐related infection (FRI) and periprosthetic joint infection (PJI). Preclinical models focusing on these areas should include all factors associated with orthopedic implant‐associated infections and consider the need for revision surgery [25]. Important FRI‐specific features to be considered for example in preclinical models include the crea‑ tion of a bone instability (fracture), soft tissue damage, and a delay in treatment (debride‑ ment and surgical fixation several hours after the accident), when mimicking an open fracture situation. The most relevant recent preclinical in vivo models available have been broadly categorized based upon their means of bacterial inoculation (direct/exogenous versus hematogenous), and whether soft tissue trauma beyond the minimum required for surgery was applied (Table 7.1) [21–23,26–38].

Direct Inoculation with Minimal Trauma Early attempts at developing in vivo models of hematogenous osteomyelitis had found that intravenous (IV) inoculation of bacteria alone in young healthy animals caused inconsistent results, and that direct inoculation of S. aureus alone into intact bone failed to create pathology mimicking chronic osteomyelitis [39]. In order to get a consistent and progressive osteomyelitis, a sclerosing agent such as sodium morrhuate (SM) can be administered to initiate local ischemia and tissue damage to the bone [39]. However, the use of sclerosing agents is now considered somewhat controversial due to the unknown effects of the sclerosing agent, which may confound results. Currently, preclinical models show reliable results with the use of direct exogenous bacterial inoculation within a surgical site involving bone damage and/or foreign body placement without additional soft tissue trauma. These models are particularly suitable

Table 7.1.  Select examples of animal models used for bone and joint infection research classified according to clinical situation. Objective of research Treatment of fraction‐related infection Prevention of fraction‐related infection

Clinical situation

Species

Location

Type of bone instability

Implant

Reference

Open fracture

Rat

Femur

Bone defect

Plate

[26]

Close fracture

Rat

Femur

Bone defect

Plate

[27]

Closed fracture

Rabbit

Tibia

Osteotomy

Intramedullary

[28]

Plate

[29]

Osteotomy

Plate

[30]

Bone defect

Plate

[32]

Tibia

Osteotomy

Intramedullary

Sheep

Tibia

Osteotomy

Plate

[35]

Open fracture

Sheep

Tibia

Osteotomy

Intramedullary

[36]

Closed fracture

Rat

Femur

Osteotomy

Intramedullary

[21]

Plate

[31]

Humerus Rat

Femur

[33]

Bone healing with infection

Pathogenesis musculoskeletal infection

Closed fracture

Mouse Rat

Femur

[34]

Osteotomies

Intramedullary

[23]

Osteotomy

Plate

[37]

Fracture (trauma)

Intramedullary

[22]

Bone defect

Plate

[38]

7  Preclinical Models of Infection in Bone and Joint Surgery

105

for research questions regarding PJI. Smeltzer et al. [40] showed that chronic osteomyelitis could be generated in an otherwise healthy rabbit, by using a devascularized segment of rabbit radius. The devascularized bone segment was inoculated with S. aureus, and it was found that the bone served as a focus for infection propagation, which did not occur to the same extent when the inoculum was added to an empty defect [40]. Equivalent early work by Andriole et  al.  [41] had demonstrated that a foreign body implanted in bone played a similar role, whereby in the presence of a piece of steel within the rabbit tibiae, an infection could develop with greater frequency than in the absence of the foreign body. Petty et  al.  [42] confirmed the high susceptibility of implants to infection. They com‑ pared different materials, and found that bone cement increased the risk for infection more than titanium or stainless steel. However, this was probably due to heat production during polymerization of polymethylmethacrylate. Recent studies showed that FRI models including bone instability (osteotomy or bone defect) followed by fixation had a reliable likelihood of infection [21,23,29–31,34,37,43–45]. Rittman and Perren [46] established an early experimental large animal model in sheep to evaluate the impact of fixation stability on healing in an infected fracture. Their prior‑ ity, reflecting clinical focus at the time, was on histological evidence of primary bone healing, and the model demonstrated that a fracture union was possible even when infected under conditions of biomechanical stability. Later studies raised concerns regarding the safety of large animal models incorporating contaminated fractures, spe‑ cifically when using intramedullary nails [36]. More recently, a sheep infection model for long‐bone plate osteosynthesis was developed and was used to evaluate the effect of a hydrophobic polycationic coating, an antimicrobial‐loaded polymer sleeve, and a vancomycin‐ modified plate on biofilm formation [35,47]. Overall, this model, based on a unilat‑ eral tibial mid‐diaphyseal osteotomy repaired with an LCP plate, showed a low incidence of complications. Small animal models have also been described. Darouiche et al. [48] created a model, which is noteworthy as one of the first in vivo models evaluating prophylactic implant coatings in the presence of a fracture. A saw osteotomy was made in rabbit tibiae, which was subsequently fixed with a chlorhexidine‐ and chloroxylenol‐coated, intramedullary Kirschner wire and inoculated with S aureus. After six weeks, the coated nail group was significantly less likely to have microbiological evidence of implant‐related infection than the uncoated control group. Worlock et al. [49] were among the first to reliably create chronic osteomyelitis in a rabbit tibia osteotomy model. They used this model to show the impact of fixation stability on infection susceptibility. The limitation of their experi‑ mental design was the unstable construct using a Kirschner wire in the tibia without interlocking bolts. More stable fixation options were later investigated in rabbits; and recent studies using intramedullary nails and plate fixation to repair bone instabilities following osteotomies of the humerus and tibia were successfully used as FRI mod‑ els [29,44,45,50]. Osteotomy models are also widely used in rodents. Plates are success‑ fully applied following femoral osteotomies in mice to characterize the impact of S. aureus in canaliculi of cortical bone and evaluate the immune response of fracture fixation with and without S. aureus infection [37,43]. A similar osteotomy model was used in rats to evaluate the necessity to remove implants following FRI, different regimens of antimi‑ crobials, and use of cationic steroid antibiotic to prevent nonunion of infected fractures [21,23,30,34]. Rat femur models with infected large cortical bone defects have also been created, allowing the evaluation of osteo‐inductive agents and osteo‐conductive scaffold materials

106

Bone and Joint Infections

in the context of osteomyelitis as well as efficacy of various delivery methods of antimicrobial in contaminated fractures [27,32,33,38]. Although stabilizing these defects represented a major challenge in the past, the recent availability of small‐scale locked plates and inter‑ locking nails have made rodents one of the main animal species used in bone and joint infection models. The pathogen to be inoculated is another important factor to consider in preclinical studies. S. aureus is by far the most inoculated pathogen in animal models of FRIs at 97% [3]. In clinical reality, numerous other pathogens cause FRI; however, these other species are currently not represented in the scientific literature [3]. Systemic antibiotics are a cornerstone in both prevention and treatment of osteomyeli‑ tis. Animal models have been used predominantly to confirm their efficacy, tailor regi‑ mens, and characterize the pharmacological parameters involved. For example [27,34], the efficacy of rifampicin, a crucial antibiotic in the medical treatment of implant‐related bone infection, was described in a subcutaneous tissue cage model using guinea pigs [51]. Because of rapid emergence of resistance when rifampicin is used as a single agent, a combination antibiotic regimen for MRSA was evaluated in this model. The data showed that daptomycin and levofloxacin are particularly effective combination partners that are able to prevent the emergence of rifampin resistance. It should be noted that subcutane‑ ous tissue cages are commonly used as a foreign body infection model. However, this animal model does not consider the special case of bone infection. Nevertheless, it pro‑ vides preclinical data that facilitate extrapolation to implant‐related bone infections. Schwank et al. [58] performed an interesting approach to antibiotic therapy in preclini‑ cal testing. The novel approach involved growing bacterial biofilms on small glass beads in vitro and exposing them to antibiotic concentrations based on normal human pharma‑ cokinetics. The authors were able to identify antibiotic combinations that could be shown to result in eradication of biofilm in vitro, and after replicating these scenarios in the guinea pig, there was a correlation between the regimens found to work in vitro with clinical outcome in biofilm infections in vivo. Numerous animal studies over the last two decades have also investigated the efficacy of local antibiotic delivery vehicles. Technology explored using in vivo models for local delivery of antiseptic or antibiotic agents including collagen sheets [52], calcium phos‑ phate pellets  [53], polysaccharide (chitosan)  [54], cross‐linked high amylose starch implants [55], biodegradable polymer beads [56], implant coatings [45], covalently bonded antibiotics [57], and gel or thermo‐responsive vehicles [27,29]. In this situation, the in vivo models used are designed to analyze outcomes such as drug release profiles, biocompat‑ ibility, effects on fracture healing, infection susceptibility, and drug resistance. A rat model designed by Lucke et al. [58] to evaluate a gentamicin‐impregnated poly‐d,l‐ lactic acid (PDLLA)–coated nail, which has since been approved for clinical use, provides a good example of translational research. Kirschner wires were coated with a PDLLA polymer containing gentamicin and implanted in the tibial medullary canal along with S. aureus. There was a significant reduction in clinical symptoms in the treatment group compared with controls. In follow‐up experiments, using the same rat model, the antibi‑ otic burst release profile and osseous drug concentrations at progressive time points were described. This study was conducted in parallel with a prospective clinical trial of the nail in open tibial fractures [59]. This particular antibiotic‐containing medical device is only one of many currently being developed, but this series of studies demonstrates the inte‑ gral role of the animal model in proving the efficacy of the coating prior to successful introduction into the clinic.

7  Preclinical Models of Infection in Bone and Joint Surgery

107

Figure 7.2.  Mouse infection model with bone instability and locking plate placed during a first surgical procedure with bacterial inoculation that underwent a second surgical procedure for debridement of surgical site that showed an abscess (arrow).

Successfully treating or managing an established active bone or joint infection is a distinctly more challenging undertaking than preventing infection. Biofilm formation, poor osse‑ ous perfusion of antibiotics, and abscess formation necessitate a multimodal medical and surgical approach. In vivo models for treatment of infection are correspondingly more complex. At least two surgical procedures are required – one to inoculate, and the second to treat, often requiring surgical site debridement (Figure 7.2), with an intervening period to allow infection to develop. Interventions effective in prophylaxis will not necessarily be effective in treatment. For example, using a canine bone infection model, it was shown that PMMA loaded with gentamicin was successful in preventing the development of an infection. However, it was unable to successfully treat an active infection [60]. The clinical conditions surrounding the treatment of an active infection, including biofilm formation, intracellular bacteria, and tissue necrosis, represent a significantly more challenging ­target for any antibiotic‐loaded biomaterial.

108

Bone and Joint Infections

Animal Models of Bone and Joint Infection Incorporating Trauma Infection following open fractures causes significant clinical morbidity and creates unique management challenges [25]. Soft tissue injury in trauma patients is known to increase the FRI rate [61] and to significantly increase infection rates after a standardized closed soft tissue injury in rats [62]. However, traumatic fractures and soft tissue injuries add an additional level of complexity to infection models and are incorporated in only 25% of studies, primarily due to the burden upon the experimental animal [3]. As described ear‑ lier, most researchers performed an osteotomy or bone defect to simulate a fracture. Although creating a traumatic (“true”) fracture is more realistic, creating an osteotomy decreases confounding factors from fracture configuration, improves reproducibility of in vivo models, and keeps the number of animals to a minimum. Furthermore, fracture stability can be more difficult to achieve in traumatic fractures compared to a controlled osteotomy. Regardless, creating a real fracture is far more realistic and more comparable to clinical situations where fractures are often accompanied by soft tissue damage, hema‑ tomas, and potential vascular compromise. Techniques used to create a fracture associ‑ ated with soft tissue trauma in preclinical models include blunt trauma from a weight being dropped on the bone [63–66], three‐point bending apparatus to mimic traumatic forces  [67], firing a steel fragment to the tibia  [68] or with a haemostat  [69]. Lindsey et al. [63] reported on a rat femur model, where they were able to create a reproducible closed fracture with a blunt impact guillotine, fix it with a Kirschner wire, and inoculate S. aureus to create local infection without overwhelming sepsis. This model was later used to demonstrate the importance of timing of antibiotic administration and surgery regard‑ ing the rate of infection in wounds contaminated with S. aureus. By varying the timing of both interventions, it was shown that early antibiotic therapy was the single most impor‑ tant factor for the risk of infection [70]. A delay in surgery did result in an increase in infection rate, though there was no significant increase between a delay of 6 and 24 h. Only one preclinical model of FRI including all clinical features of the condition (frac‑ ture creation, soft tissue damage, delay in treatment) was described. This model uses Sprague‐Dawley rats and creates a femoral fracture using a blunt trauma from a weight dropped from a height. Following fracture creation, the surgical site is inoculated with a S. aureus bacterial suspension and left open for 1h prior to surgical stabilization using an intramedullary K‐wire [63‑66]. The potential ethical issues and complications inherent in these models are exemplified in a recent study, where efforts were made to reliably pro‑ duce a traumatic tibial fracture in a rat model  [71]. Subject animals were divided into small groups depending on fracture configuration and whether the fracture was fixed with an intramedullary pin or with external coaptation. The authors reported extremity necrosis at day 7 postoperatively and several animals showing weight loss and signs of obvious malnutrition.

Hematogenous Models Hematogenous osteomyelitis is a particular issue in pediatric medicine where septic arthritis and infection in the adjacent metaphyses of long bones is relatively common [72]. It is also a common cause for late infection in previously well‐functioning prosthetic joints, with a 39% risk of PJI following a S. aureus bacteremia [73]. Acknowledging this distinct etiology, a number of authors have attempted to create models of hematogenous

7  Preclinical Models of Infection in Bone and Joint Surgery

109

osteomyelitis using S. aureus, and a focal bone lesion typically created either concurrently or prior to inoculation. An early model, created by Deysine et al. [74], involved injecting a combination of barium and 5 × 105 CFU of S. aureus simultaneously into the tibial nutrient artery of dogs. The authors reliably created acute osteomyelitis, but there was an unacceptably high mortality rate from sepsis. Other similar models in chickens [75] and rabbits [76] encountered the same problem, describing narrow safety margins in inocu‑ lum dose when administered systemically. Hienz et al. [77], however, were able to create a model using rats with no mortality during the 14‐day study period. They injected SM locally into the mandible and tibia and inoculated varying doses of S. aureus intrave‑ nously into the tail vein to determine the dose required to infect 50% (ID50) and 100% (ID100) of animals. Animals inoculated with bacteria, but spared the focal SM injection, did not develop osteomyelitis, again demonstrating the importance of local perturbation of bone physiology. An alternative approach was taken by Whalen et al. [78] who designed a model to mimic pediatric hematogenous osteomyelitis. Using skeletally immature rab‑ bits, they created a partial growth plate fracture at the proximal tibial metaphysis and administered S. aureus intravenously via an ear vein, reliably causing focal acute osteo‑ myelitis without metastatic infection. Skeletally immature (20–24 weeks) rabbits were used in this study, in order to simulate pediatric hematogenous osteomyelitis. The impor‑ tance of choosing a model replicating clinical cases was highlighted by a study using a model of localized osteomyelitis by injecting S. aureus unilaterally into the femoral artery of domestic pigs [79]. The use of juvenile domestic pigs (40 kg) led to early euthanasia of several animals due to lameness, shallow respiration, fever, and anorexia in contrast to fewer complications and euthanasia in younger pigs (20 kg). In addition, osteomyelitis caused by S. aureus was seen in six of the seven 20 kg pigs and in none of the 40 kg pigs [80]. More recently, a murine hematogenous osteomyelitis model that closely mimics the human infection was described [81,82]. The mice were infected intravenously with 106 CFUs of S. aureus via the tail vein and subsequently developed chronic osteomyelitis.

Future Directions Animal models of bone and joint infection are developed with the primary aim of improving outcomes in clinical medicine. In theory, they allow in vivo evaluation of potential therapies, prophylactics, and diagnostics without the costs, safety, and ethical issues associated with human clinical trials. Prophylactic strategies in orthopedic surgery, such as systemic and local antibiotic therapy, have evolved over the last four decades. Clinical use of IV antibiotics and antibi‑ otic‐loaded bone cement in arthroplasty grew sporadically throughout the 1970s and was followed only later with detailed characterization in controlled animal models  [83]. Mainstay surgical techniques in the treatment of established infection such as debride‑ ment, stabilization, and lavage have also been characterized after the fact, rather than developed with the help of animal models. The limitations involved with these traditional strategies, and an ever‐expanding understanding of bacterial virulence factors, have, however, fueled significant efforts to develop new technology to combat osteomyelitis, which inevitably increases the demand for reliable and valid in vivo models [84]. Promising new approaches to prevent infection include (i) modification of implant surface to deter bacterial adhesion [57]; (ii) coating implants with degradable polymers that elute high concentrations of antibiotic into the local milieu (without causing toxic

110

Bone and Joint Infections

systemic concentrations) [58,85]; (iii) new drugs targeted at either the genes or effector molecules of adhesion [86], quorum sensing [87] or RNA processing [88]; and (iv) vaccine development against biofilm‐forming bacteria [89].

Conclusions Animal models in modern biomedical research are indispensable in the development of novel interventional and diagnostic technologies. The success of any future anti‐infective technology will depend upon proper evaluation in appropriate animal models. Robust assessment of the performance of any clinical device may require testing in comparatively low‐burden animal models, although higher‐burden models, including, for example, frac‑ ture creation and localized tissue damage, will be required in certain cases. The develop‑ ment of refined small animal models will enable screening of candidate technologies that are gated at an early stage to reduce the need for more burdensome investigations of any but the most promising candidates. Real‐time, in vivo estimation of bacterial burden is also likely to be a key area for a reduction in the number of animals required in the future. Preclinical models investigating musculoskeletal infection should recapitulate specific features of the clinical condition [2,26]. Current preclinical in vivo musculoskeletal infec‑ tion models rarely mimic a clinical situation of musculoskeletal infection. In addition to the overrepresentation of S. aureus in infection models, other clinically important factors such as soft tissue trauma are rarely recreated. Refinement of existing preclinical models and development of new models more resembling to clinical scenarios are needed to make progress in the field of prevention and treatment of bone and joint infection. Furthermore, the recent development of humanized are a promising avenue to increase translatability of preclinical research using murine model of infection [13].

Key Points ●●

●●

●●

●●

●●

Different animal species may vary in bone structure, susceptibility to infection, adaptive immune response to bacteria, and specificity of bacterial toxins. The implant systems available for laboratory animals are improving, with biomechan‑ ically defined fracture fixation now available in rodent models. Investigators should weigh the importance of clinical relevance versus burden upon the animal when deciding upon the particular model chosen. Testing novel prophylactic measures requires different models in comparison with testing of novel treatments for bone and joint infection. The use of immunodeficient mice reconstituted with a human hematopoietic immune system allows assessment of human‐specific toxins in murine model of infection.

References 1. Sabaté Brescó M, O’Mahony L, Zeiter S, et al. Influence of fracture stability on Staphylococcus epidermidis and Staphylococcus aureus infection in a murine femoral fracture model. Eur Cell Mater. 2017;34:321–340.

7  Preclinical Models of Infection in Bone and Joint Surgery

111

2. Moriarty TF, Harris LG, Mooney RA, et  al. Recommendations for design and conduct of preclinical in vivo studies of orthopedic device‐related infection. Journal of Orthopaedic Research. 2019;37(2):271–287. 3. Vanvelk N, Morgenstern M, Moriarty TF, et al. Preclinical in vivo models of fracture‐related infection: a systematic review and critical appraisal. Eur Cell Mater. 2018;36:184–199. 4. Aerssens J, Boonen S, Lowet G, et al. Interspecies differences in bone composition, density, and quality: potential implications for in vivo bone research. Endocrinology. 1998;139(2): 663–670. 5. Currey JD. Bones: structure and mechanics: Princeton University Press, 2006. 6. Reinwald S, Burr D. Review of nonprimate, large animal models for osteoporosis research. J Bone Miner Res. 2008;23(9):1353–1368. 7. Torbert JT, Joshi M, Moraff A, et al. Current bacterial speciation and antibiotic resistance in deep infections after operative fixation of fractures. J Orthop Trauma. 2015;29(1):7–17. 8. Gieling F, Peters S, Erichsen C, et  al. Bacterial osteomyelitis in veterinary orthopaedics: Pathophysiology, clinical presentation and advances in treatment across multiple species. The Veterinary Journal. 2019;250:44–54. 9. Sung JM, Lloyd DH, Lindsay JA. Staphylococcus aureus host specificity: comparative genom‑ ics of human versus animal isolates by multi‐strain microarray. Microbiology (Reading). 2008;154(Pt 7):1949–1959. 10. Löffler B, Hussain M, Grundmeier M, et al. Staphylococcus aureus panton‐valentine leukoci‑ din is a very potent cytotoxic factor for human neutrophils. PLoS Pathog. 2010;6(1):e1000715. 11. Gillet Y, Issartel B, Vanhems P, et al. Association between Staphylococcus aureus strains carry‑ ing gene for Panton‐Valentine leukocidin and highly lethal necrotising pneumonia in young immunocompetent patients. Lancet. 2002;359(9308):753–759. 12. Shallcross LJ, Fragaszy E, Johnson AM, et  al. The role of the Panton‐Valentine leucocidin toxin in staphylococcal disease: a systematic review and meta‐analysis. Lancet Infect Dis. 2013;13(1):43–54. 13. Prince A, Wang H, Kitur K, et al. Humanized mice exhibit increased susceptibility to Staphylococcus aureus pneumonia. The Journal of Infectious Diseases. 2016;215(9):1386–1395. 14. Lipinska U, Hermans K, Meulemans L, et al. Panton‐Valentine leukocidin does play a role in the early stage of Staphylococcus aureus skin infections: a rabbit model. PLoS One. 2011; 6(8):e22864. 15. Spaan AN, Henry T, van Rooijen WJM, et  al. The staphylococcal toxin Panton‐Valentine Leukocidin targets human C5a receptors. Cell Host Microbe. 2013;13(5):584–594. 16. Thammavongsa V, Kim HK, Missiakas D, et al. Staphylococcal manipulation of host immune responses. Nat Rev Microbiol. 2015;13(9):529–543. 17. Yong KSM, Her Z, Chen Q. Humanized mice as unique tools for human‐specific studies. Arch Immunol Ther Exp (Warsz). 2018;66(4):245–266. 18. Shultz LD, Brehm MA, Garcia‐Martinez JV, et al. Humanized mice for immune system inves‑ tigation: progress, promise and challenges. Nat Rev Immunol. 2012;12(11):786–798. 19. Shi Y, Inoue H, Wu JC, et al. Induced pluripotent stem cell technology: a decade of progress. Nat Rev Drug Discov. 2017;16(2):115–130. 20. Claes L, Meyers N, Schülke J, et  al. The mode of interfragmentary movement affects bone formation and revascularization after callus distraction. PLOS ONE. 2018;13(8):e0202702. 21. Bilgili F, Balci HI, Karaytug K, et  al. Can normal fracture healing be achieved when the implant is retained on the basis of infection? An experimental animal model. Clin Orthop Relat Res. 2015;473(10):3190–3196. 22. Gilbert SR, Camara J, Camara R, et al. Contaminated open fracture and crush injury: a murine model. Bone Res. 2015;3:14050–14050. 23. Schindeler A, Yu NY, Cheng TL, et al. Local delivery of the cationic steroid antibiotic CSA‐90 enables osseous union in a rat open fracture model of Staphylococcus aureus infection. J Bone Joint Surg Am. 2015;97(4):302–309.

112

Bone and Joint Infections

24. Histing T, Garcia P, Matthys R, et  al. An internal locking plate to study intramembranous bone healing in a mouse femur fracture model. J Orthop Res. 2010;28(3):397–402. 25. Depypere M, Morgenstern M, Kuehl R, et  al. Pathogenesis and management of fracture‐ related infection. Clinical Microbiology and Infection. 2020;26(5):572–578. 26. Brown KV, Penn‐Barwell JG, Rand BC, et al. Translational research to improve the treatment of severe extremity injuries. Journal of the Royal Army Medical Corps. 2014;160(2): 167–170. 27. Rand BC, Penn‐Barwell JG, Wenke JC. Combined local and systemic antibiotic delivery improves eradication of wound contamination: An animal experimental model of contami‑ nated fracture. Bone Joint J. 2015;97‐b(10):1423–1427. 28. Li Y, Chen S‐K, Li L, et al. Bone defect animal models for testing efficacy of bone substitute biomaterials. Journal of Orthopaedic Translation. 2015;3(3):95–104. 29. Ter Boo GA, Arens D, Metsemakers WJ, et al. Injectable gentamicin‐loaded thermo‐respon‑ sive hyaluronic acid derivative prevents infection in a rabbit model. Acta Biomater. 2016;43:185–194. 30. Lovati AB, Drago L, Bottagisio M, et al. Systemic and local administration of antimicrobial and cell therapies to prevent methicillin‐resistant Staphylococcus epidermidis‐induced femoral nonunions in a rat model. Mediators Inflamm. 2016;2016:9595706. 31. Lovati AB, Romanò CL, Bottagisio M, et  al. Modeling Staphylococcus epidermidis‐ induced non‐unions: Subclinical and clinical evidence in rats. PLoS One. 2016;11(1): e0147447. 32. Penn‐Barwell JG, Baker B, Wenke JC. Local bismuth thiols potentiate antibiotics and reduce infection in a contaminated open fracture model. J Orthop Trauma. 2015;29(2):e73–78. 33. Tennent DJ, Shiels SM, Sanchez CJ, Jr., et al. Time‐dependent effectiveness of locally applied vancomycin powder in a contaminated traumatic orthopaedic wound model. J Orthop Trauma. 2016;30(10):531–537. 34. Sethi S, Thormann U, Sommer U, et al. Impact of prophylactic CpG Oligodeoxynucleotide applica‑ tion on implant‐associated Staphylococcus aureus bone infection. Bone. 2015;78:194–202. 35. Schaer TP, Stewart S, Hsu BB, et al. Hydrophobic polycationic coatings that inhibit biofilms and support bone healing during infection. Biomaterials. 2012;33(5):1245–1254. 36. Hill PF, Clasper JC, Parker SJ, et al. Early intramedullary nailing in an animal model of a heav‑ ily contaminated fracture of the tibia. J Orthop Res. 2002;20(4):648–653. 37. de Mesy Bentley KL, Trombetta R, Nishitani K, et  al. Evidence of Staphylococcus aureus deformation, proliferation, and migration in canaliculi of live cortical bone in murine models of osteomyelitis. J Bone Miner Res. 2017;32(5):985–990. 38. Seebach E, Holschbach J, Buchta N, et al. Mesenchymal stromal cell implantation for stimula‑ tion of long bone healing aggravates Staphylococcus aureus induced osteomyelitis. Acta Biomater. 2015;21:165–177. 39. Schrman L, Janota M, Lewin P. The production of experimental osteomyelitis: preliminary report. Journal of the American Medical Association. 1941;117(18):1525–1529. 40. Smeltzer MS, Thomas JR, Hickmon SG, et al. Characterization of a rabbit model of staphylo‑ coccal osteomyelitis. J Orthop Res. 1997;15(3):414–421. 41. Andriole VT, Nagel DA, Southwick WO. A paradigm for human chronic osteomyelitis. J Bone Joint Surg Am. 1973;55(7):1511–1515. 42. Petty W, Spanier S, Shuster JJ, et al. The influence of skeletal implants on incidence of infec‑ tion. Experiments in a canine model. J Bone Joint Surg Am. 1985;67(8):1236–1244. 43. Rochford ETJ, Sabaté Brescó M, Zeiter S, et  al. Monitoring immune responses in a mouse model of fracture fixation with and without Staphylococcus aureus osteomyelitis. Bone. 2016;83:82–92. 44. Arens D, Wilke M, Calabro L, et al. A rabbit humerus model of plating and nailing osteosyn‑ thesis with and without Staphylococcus aureus osteomyelitis. Eur Cell Mater. 2015;30:148–161; discussion 61–62.

7  Preclinical Models of Infection in Bone and Joint Surgery

113

45. Prinz C, Elhensheri M, Rychly J, et al. Antimicrobial and bone‐forming activity of a copper coated implant in a rabbit model. J Biomater Appl. 2017;32(2):139–149. 46. Rittmann W, Perren S. Cortical bone healing after internal fixation and fracture. Springer/ Verlag. Berlin/Heidelberg/New York, 1974. 47. Stewart S, Barr S, Engiles J, et al. Vancomycin‐modified implant surface inhibits biofilm forma‑ tion and supports bone‐healing in an infected osteotomy model in sheep: a proof‐of‐concept study. The Journal of bone and joint surgery American volume. 2012;94(15):1406–1415. 48. Darouiche RO, Farmer J, Chaput C, et al. Anti‐infective efficacy of antiseptic‐coated intramed‑ ullary nails. J Bone Joint Surg Am. 1998;80(9):1336–1340. 49. Worlock P, Slack R, Harvey L, et al. The prevention of infection in open fractures. An experi‑ mental study of the effect of antibiotic therapy. J Bone Joint Surg Am. 1988;70(9):1341–1347. 50. Metsemakers WJ, Schmid T, Zeiter S, et al. Titanium and steel fracture fixation plates with dif‑ ferent surface topographies: Influence on infection rate in a rabbit fracture model. Injury. 2016;47(3):633–639. 51. John AK, Baldoni D, Haschke M, et al. Efficacy of daptomycin in implant‐associated infection due to methicillin‐resistant Staphylococcus aureus: importance of combination with rifampin. Antimicrob Agents Chemother. 2009;53(7):2719–2724. 52. Riegels‐Nielsen P, Espersen F, Hölmich LR, et al. Collagen with gentamicin for prophylaxis of postoperative infection. Staphylococcus aureus osteomyelitis studied in rabbits. Acta Orthop Scand. 1995;66(1):69–72. 53. Beardmore AA, Brooks DE, Wenke JC, et al. Effectiveness of local antibiotic delivery with an osteoinductive and osteoconductive bone‐graft substitute. J Bone Joint Surg Am. 2005; 87(1):107–112. 54. Paiva Costa L, Moreira Teixeira LE, Maranhão Lima GS, et  al. Effectiveness of Chitosan Films Impregnated With Ciprofloxacin for the Prophylaxis of Osteomyelitis in Open Fractures: An Experimental Study in Rats. Arch Trauma Res. 2016;5(3):e36952–e52. 55. Huneault LM, Lussier B, Dubreuil P, et al. Prevention and treatment of experimental osteomy‑ elitis in dogs with ciprofloxacin‐loaded crosslinked high amylose starch implants. J Orthop Res. 2004;22(6):1351–1357. 56. Ambrose CG, Clyburn TA, Louden K, et al. Effective treatment of osteomyelitis with biode‑ gradable microspheres in a rabbit model. Clin Orthop Relat Res. 2004. doi: 10.1097/01. blo.0000126303.41711.a2(421):293‐9. 57. Antoci V, Jr., Adams CS, Parvizi J, et al. The inhibition of Staphylococcus epidermidis biofilm formation by vancomycin‐modified titanium alloy and implications for the treatment of periprosthetic infection. Biomaterials. 2008;29(35):4684–4690. 58. Lucke M, Schmidmaier G, Sadoni S, et al. Gentamicin coating of metallic implants reduces implant‐related osteomyelitis in rats. Bone. 2003;32(5):521–531. 59. Fuchs T, Stange R, Schmidmaier G, et al. The use of gentamicin‐coated nails in the tibia: pre‑ liminary results of a prospective study. Arch Orthop Trauma Surg. 2011;131(10): 1419–1425. 60. Fitzgerald RH, Jr. Experimental osteomyelitis: description of a canine model and the role of depot administration of antibiotics in the prevention and treatment of sepsis. J Bone Joint Surg Am. 1983;65(3):371–380. 61. Young K, Aquilina A, Chesser TJS, et  al. Open tibial fractures in major trauma centres: A national prospective cohort study of current practice. Injury. 2019;50(2):497–502. 62. Kälicke T, Schlegel U, Printzen G, et al. Influence of a standardized closed soft tissue trauma on resistance to local infection. An experimental study in rats. J Orthop Res. 2003;21(2): 373–378. 63. Lindsey BA, Clovis NB, Smith ES, et al. An animal model for open femur fracture and osteo‑ myelitis: Part I. J Orthop Res. 2010;28(1):38–42. 64. Boyce BM, Lindsey BA, Clovis NB, et al. Additive effects of exogenous IL‐12 supplementation and antibiotic treatment in infection prophylaxis. J Orthop Res. 2012;30(2):196–202.

114

Bone and Joint Infections

65. Li B, Jiang B, Boyce BM, et al. Multilayer polypeptide nanoscale coatings incorporating IL‐12 for the prevention of biomedical device‐associated infections. Biomaterials. 2009;30(13): 2552–2558. 66. Lindsey BA, Clovis NB, Smith ES, et al. An animal model for open femur fracture and osteo‑ myelitis—Part II: Immunomodulation with systemic IL‐12. Journal of Orthopaedic Research. 2010;28(1):43–47. 67. Ashhurst DE, Hogg J, Perren SM. A method for making reproducible experimental fractures of the rabbit tibia. Injury. 1982;14(3):236–242. 68. Hill PF, Watkins PE. The prevention of experimental osteomyelitis in a model of gunshot fracture in the pig. European Journal of Orthopaedic Surgery & Traumatology. 2001;11(4): 237–241. 69. Petri WH, 3rd, Schaberg SJ. The effects of antibiotic‐supplemented bone allografts on con‑ taminated, partially avulsive fractures of the canine ulna. J Oral Maxillofac Surg. 1984; 42(11):699–704. 70. Penn‐Barwell JG, Murray CK, Wenke JC. Early antibiotics and debridement independently reduce infection in an open fracture model. J Bone Joint Surg Br. 2012;94(1):107–112. 71. Shi E, Chen G, Qin B, et al. A novel rat model of tibial fracture for trauma researches: a com‑ bination of different types of fractures and soft tissue injuries. J Orthop Surg Res. 2019;14(1):333. 72. Brown DW, Sheffer BW. Pediatric septic arthritis: An update. Orthop Clin North Am. 2019;50(4):461–470. 73. Sendi P, Banderet F, Graber P, et al. Periprosthetic joint infection following Staphylococcus aureus bacteremia. J Infect. 2011;63(1):17–22. 74. Deysine M, Rosario E, Isenberg HD. Acute hematogenous osteomyelitis: an experimental model. Surgery. 1976;79(1):97–99. 75. Emslie KR, Nade S. Pathogenesis and treatment of acute hematogenous osteomyelitis: evaluation of current views with reference to an animal model. Rev Infect Dis. 1986;8(6): 841–849. 76. Poultsides LA, Papatheodorou LK, Karachalios TS, et al. Novel model for studying hematog‑ enous infection in an experimental setting of implant‐related infection by a community‐ acquired methicillin‐resistant S. aureus strain. J Orthop Res. 2008;26(10):1355–1362. 77. Hienz SA, Sakamoto H, Flock JI, et al. Development and characterization of a new model of hematogenous osteomyelitis in the rat. J Infect Dis. 1995;171(5):1230–1236. 78. Whalen JL, Fitzgerald RH, Jr., Morrissy RT. A histological study of acute hematogenous osteomyelitis following physeal injuries in rabbits. J Bone Joint Surg Am. 1988;70(9): 1383–1392. 79. Nielsen OL, Afzelius P, Bender D, et al. Comparison of autologous (111)In‐leukocytes, (18) F‐FDG, (11)C‐methionine, (11)C‐PK11195 and (68)Ga‐citrate for diagnostic nuclear imaging in a juvenile porcine haematogenous staphylococcus aureus osteomyelitis model. Am J Nucl Med Mol Imaging. 2015;5(2):169–182. 80. Alstrup AKO, Nielsen KM, Schønheyder HC, et al. Refinement of a hematogenous localized osteomyelitis model in pigs. Scand J Lab Anim Sci. 2016;42:1–4. 81. Horst SA, Hoerr V, Beineke A, et al. A novel mouse model of Staphylococcus aureus chronic osteomyelitis that closely mimics the human infection: an integrated view of disease pathogen‑ esis. Am J Pathol. 2012;181(4):1206–1214. 82. Tuchscherr L, Geraci J, Löffler B. Staphylococcus aureus Regulator Sigma B is Important to Develop Chronic Infections in Hematogenous Murine Osteomyelitis Model. Pathogens. 2017;6(3):31. 83. Norden CW. Experimental osteomyelitis. IV. Therapeutic trials with rifampin alone and in combination with gentamicin, sisomicin, and cephalothin. J Infect Dis. 1975;132(5): 493–499. 84. Montanaro L, Campoccia D, Arciola CR. Advancements in molecular epidemiology of implant infections and future perspectives. Biomaterials. 2007;28(34):5155–5168. 85. Xue‐Wen H, Jie P, Xian‐Yuan A, et al. Inhibition of bacterial translation and growth by pep‑ tide nucleic acids targeted to domain II of 23S rRNA. J Pept Sci. 2007;13(4):220–226.

7  Preclinical Models of Infection in Bone and Joint Surgery

115

86. Hatamoto M, Ohashi A, Imachi H. Peptide nucleic acids (PNAs) antisense effect to bacterial growth and their application potentiality in biotechnology. Appl Microbiol Biotechnol. 2010;86(2):397–402. 87. Teiber JF, Horke S, Haines DC, et al. Dominant role of paraoxonases in inactivation of the Pseudomonas aeruginosa quorum‐sensing signal N‐(3‐oxododecanoyl)‐L‐homoserine lactone. Infect Immun. 2008;76(6):2512–2519. 88. Eidem TM, Roux CM, Dunman PM. RNA decay: a novel therapeutic target in bacteria. Wiley Interdiscip Rev RNA. 2012;3(3):443–454. 89. McCarthy AJ, Lindsay JA. Genetic variation in Staphylococcus aureus surface and immune evasion genes is lineage associated: implications for vaccine design and host‐pathogen interac‑ tions. BMC Microbiol. 2010;10:173.

Chapter 8

Native Joint Arthritis in Children Pablo Yagupsky

Introduction Bacterial and fungal joint infections in children are medical emergencies. If diagnosis and treatment are delayed or inadequate, severe morbidity, irreversible joint damage, and even fatalities may result. Septic arthritis is more common in childhood than in any other period, and more than half of cases are diagnosed in individuals younger than 20 years of age.

Epidemiology The estimated annual incidence of pediatric joint infections in the western world ranges between 2 and 10 cases per 100,000 and is much higher in the developing world and indigent populations [1,2]. A male‐to‐female ratio >1 has been consistently reported in large patients’ series [3]. Because over 95% of cases of pediatric septic arthritis are of hematogenous origin, the age distribution of patients with joint infection is markedly skewed to the left, reflecting the increased attack rate of bacteremia in early childhood  [3]. In a large case series comprising 725 children, 54% percent of all cases were diagnosed below the age of 2 years, 25% in children aged 2–5 years, 15% among children aged 6–10 years, and the remaining 6% in the 11–15‐year‐old group [4]. The delayed maturation of the T‐cell independent arm of the immune system in humans results in impaired production of antibodies to bacterial polysaccharides below 2–4 years of age  [5]. Thus, this age group has an increased susceptibility to encapsulated organisms such as Haemophilus influenzae type b, Kingella kingae, or pneumococci [5,6]. On the other hand, the incidence of infectious arthritis in infants younger than 6 months is low, indicating vertically acquired immunity and relative lack of social contacts, resulting in reduced exposure to potential pathogens in early life.

Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

117

118

Bone and Joint Infections

Microbiology Specific Microorganisms and Predisposition Because joint infections in children usually result from hematogenous seeding, the etiology of septic arthritis frequently overlaps that of pediatric bacteremia. However, some microorganisms such as Staphylococcus aureus or K. kingae are remarkably overrepresented in childhood arthritis, indicating joint tissue tropism [7]. As a rule, pediatric septic arthritis is caused by the intra‐articular invasion of a single bacterial (or fungal) species. Isolation of multiple organisms should raise the suspicion of culture contamination, immunodeficiency, intravenous drug use, or penetrating trauma with direct inoculation of microorganisms into the joint space. The patient’s age and presence of associated extra‐articular symptoms and signs may provide clues to the likely bacterial etiology, as shown in Tables 8.1 and 8.2. Staphylococcus aureus is the most common cause of joint infections in neonates, as well as in children older than 4 years. This organism is characterized by a wide array of virulence factors and genetic determinants of antibiotic resistance. In recent years, methicillin‐resistant strains of S. aureus (MRSA) are being increasingly detected in many regions, whereas the rate of infection caused by methicillin‐susceptible S. aureus remains Table 8.1.  Etiology of pediatric hematogenous septic arthritis by age group. Age

Organism

15 years

Staphylococcus aureus Neisseria gonorrhoeaec

 In premature babies with indwelling vascular catheters.  Among unvaccinated and incompletely vaccinated children. c  In sexually active adolescents.

a

b

8  Native Joint Arthritis in Children

Table 8.2.  Etiology of pediatric septic arthritis and associated clinical conditions. Associated Condition

Possible etiology

Skin Pyoderma

Staphylococcus aureus, Streptococcus pyogenes

Varicella lesions

Streptococcus pyogenes, Kingella kingae

Erythematous rash

Streptococcus pyogenes

Erythema migrans

Borrelia burgdorferi

Petechial rash

Neisseria meningitidis

Mucosae Gingivostomatitis

Kingella kingae

Urethritis/genital infection

Neisseria gonorrhoeae, Ureaplasma spp., Mycoplasma hominis

Respiratory system Pneumonia

Staphylococcus aureus, Streptococcus pneumoniae, Haemophilus influenzae type b

Necrotizing pneumonia

Staphylococcus aureus (especially community‐acquired MRSA)

Cardiovascular Endocarditis

Staphylococcus aureus, Kingella kingae

Central nervous system Meningitis

Streptococcus agalactiae, Streptococcus pneumoniae, Haemophilus influenzae type b, Neisseria meningitidis, Borrelia burgdorferi

Hepatosplenomegaly

Brucella spp.

Multifocal involvement

Staphylococcus aureus, Haemophilus influenzae type b, Neisseria gonorrhoeae, Brucella spp.

Hemoglobinopathies

Salmonella enterica, Streptococcus pneumoniae, other Enterobacteriaceae

Immune system HIV infection

Staphylococcus aureus, Streptococcus pneumoniae, Mycobacterium spp., Nocardia spp., fungi

X‐linked agammaglobulinemia

Encapsulated bacteria, Mycoplasma spp. and Ureaplasma spp.

Common variable immunodeficiency

Mycoplasma spp. and Ureaplasma spp.

Chronic granulomatous disease

Staphylococcus aureus, Serratia marcescens, Pseudomonas aeruginosa, non‐tuberculous mycobacteria, Aspergillus spp.

Addictions Intravenous‐drug addicts

Staphylococcus aureus, Pseudomonas aeruginosa

119

120

Bone and Joint Infections

stable [8]. Community‐associated MRSA (CA‐MRSA) infections affect patients lacking traditional risk factors for nosocomial MRSA disease; most strains elaborate the Panton‐ Valentine leukocidin (PVL), that causes lysis of white blood cells, and are generally susceptible to antibiotics other than β‐lactams [8–10]. Infections with CA‐MRSA involve skin, soft tissues, the lung, and the skeletal system, and are characterized by remarkable tissue destruction, a pronounced inflammatory response, and a high incidence of complications requiring repeated surgical interventions, intensive care unit admissions, longer hospitalization, and results in residual morbidities [10]. Streptococcus pyogenes (group A Streptococcus) is isolated in 10–20% of preschool and early‐school children with septic arthritis. Most cases are detected during the winter and spring [11], and it is especially common in patients with concomitant skin infections or chickenpox [12]. Despite the exquisite susceptibility of S. pyogenes to β‐lactam antibiotics, the disease may be associated with severe sepsis, toxic shock, and mortality [12]. Streptococcus agalactiae (group B Streptococcus) is diagnosed almost exclusively in the neonatal period [13]. Remarkably, S. agalactiae arthritis commonly affects the shoulder. In children born in breech presentation, it typically involves the hip joint, suggesting that trauma and local hyperemia in the course of bacteremia facilitate seeding of the organism into the articular space. The disease usually occurs as a manifestation of late‐onset disease (7–90 days after birth). This may explain that current guidelines to detect maternal colonization and prevent early transmission of the bacterium to the offspring have not substantially reduced its incidence [14]. Streptococcus pneumoniae arthritis is most common between the ages of six months and two years  [14]. In countries, where the conjugate pneumococcal vaccine has been introduced, the incidence of invasive S. pneumoniae diseases, including those affecting the skeletal system, has substantially decreased [15]. Haemophilus influenzae type b was the most common cause of septic arthritis in children younger than 2 years before the advent of the conjugate vaccine, accounting for almost one half of the cases [16]. Children with H. influenzae type b arthritis frequently presented with other foci of infection such as meningitis (in 30% of patients), osteomyelitis (in 22%), cellulitis (in 30%), pneumonia (in 4%), and otitis media (in 35%)  [16]. Nowadays, H. influenzae type b has become rare in countries where immunization coverage is high [17], but there is a worldwide increase in the incidence of invasive disease, including skeletal system infections, caused by H. influenzae type a [18]. Arthritis due to encapsulated H. influenzae strains other than a or b, non‐encapsulated H. influenzae, and Haemophilus species other than influenzae almost exclusively affects immunocompromised children [19]. The increasing use of blood culture vials for seeding skeletal system exudates and species‐specific nucleic acid amplification tests (NAATs) has resulted in the recognition of K. kingae, a fastidious Gram‐negative member of the normal pharyngeal flora, as the most common etiology of septic arthritis and other skeletal system infections below 4 years of age [7,20,21]. Antecedent or concomitant stomatitis and/or signs of an upper respiratory tract infection are frequent, suggesting that invasion of the bloodstream and dissemination of the organism to the joints is facilitated by breaching of the mucosal layer by a preceding intercurrent viral disease. The disease is characterized by a mild systemic and local inflammatory reaction, requiring a high index of suspicion  [21]. Outbreaks of K. kingae osteoarthritis and other invasive diseases have been reported in daycare center facilities [21]. Although the incidence of arthritis in invasive meningococcal disease is as high as 14%, true invasion of the joint by Neisseria meningitidis is uncommon [22]. In most cases, signs

8  Native Joint Arthritis in Children

121

of joint inflammation develop several days after initiation of antibiotic therapy and the synovial fluid is usually sterile, suggesting an immune‐complex‐mediated phenomenon [22]. Recurrent disease, a prolonged course, isolation of uncommon N. meningitidis serogroups, and family clustering of cases should raise the possibility of complement or properdin deficiencies [23]. Neisseria gonorrhoeae becomes common in sexually active adolescents and its isolation in children beyond the neonatal period is definitive proof of sexual abuse. It may be pauciarticular and associated with tenosynovitis and skin lesions. Rarely, gonococci may infect the joint in the course of disseminated disease in neonates born to infected mothers [24]. Invasion of the joint space by Salmonella enterica has been reported in children suffering from sickle cell anemia and other hemoglobinopathies, and in those living in poverty in developing countries. Other Enterobacteriaceae, especially Escherichia coli and Klebsiella pneumoniae, are associated with suppurative arthritis in the neonatal period and immunocompromised patients [25]. Pseudomonas aeruginosa is a rare cause of septic arthritis in the general pediatric population. However, it may cause joint infection in neonates, in patients with vascular catheter‐related infection, immunodeficient children, and intravenous drug‐using adolescents [26]. As the result of effective public health measures, human brucellosis has been eradicated from most western countries. In children with arthritis residents of endemic countries and travelers returning from these regions, the possibility of brucellar arthritis should be considered (see Chapter 19). The disease is characterized by pain, limited mobility, and swelling, whereas local redness or warmth is rarely found. Brucellosis usually affects the weight‐bearing articulations, especially the hip (in half of the cases). The involvement of multiple joints is seen in one‐quarter of patients [27]. Lyme disease should be included in the differential diagnosis of children exposed to ticks in endemic areas who present with arthritis involving large joints (with the noticeable exception of the hip) [28]. Migratory arthralgia is present in 18% of children with Borrelia burgdorferi infections and frank arthritis in 10% [28]. Despite the presence of impressive joint inflammation and large effusions, children do not look ill, motion is possible, fever is absent in half of the cases, and the white blood cell (WBC) count is within normal limits [28]. Septic arthritis caused by Mycoplasma spp. and Ureaplasma species is almost exclusively detected in patients with X‐linked agammaglobulinemia, common variable immunodeficiency, or after organ transplantation [29]. Hematogenous septic arthritis caused by anaerobic organisms is exceptionally seen in children and is usually caused by a single bacterial species, generally a gram‐negative bacillus. Whenever a penetrating wound or bite is the mechanism of infection, multiple organisms, including both aerobes and anaerobes, may be isolated in the joint fluid culture [30]. Arthritis is the most common manifestation of tuberculosis in the skeletal system after Pott disease  [31]. Usually, Mycobacterium tuberculosis bacilli are seeded in synovial tissues by the hematogenous route during primary infection. More rarely, the disease spreads from a contiguous focus such as the invasion of the atlantoaxial joint from an apical pulmonary infection. Tubercular arthritis is monoarticular in 90% of cases and, although it can affect virtually any joint, usually involves the hip or knee  [31]. Constitutional symptoms, such as fever and weight loss, occur in only a minority of children. Granulomatous changes and cartilage erosion result in a chronic effusion and progressive joint destruction. Signs of acute inflammation are frequently absent, whereas local deformity and restricted motion range are typically observed.

122

Bone and Joint Infections

Rat‐bite fever is a rare zoonosis caused by two members of rodents’ oral flora Streptobacillus moniliformis, mostly in western countries and Australia, and by Spirillum minus in Asia [32]. The site of inoculation of the disease usually heals before a septicemic disease, frequently characterized by fever and rash, develops. Arthritis involving multiple joints is commonly seen in rat‐bite disease caused by S. moniliformis, but is rare in spirillar infection. The culture of the synovial fluid exudate is frequently negative, suggesting a reactive mechanism [32]. Candida species and coagulase‐negative staphylococci are pathogens of low‐virulence that can cause infectious arthritis in premature babies and neonates in the intensive care setting and in young infants with indwelling vascular catheters [33].

Culture‐Negative Septic Arthritis On average, in 33% of children with presumptive joint infections, blood and synovial fluid cultures reveal no growth [4,6,7], with percentages ranging between 16% [34] and 69% [35]. This wide variation reflects differences in the sensitivity of the microbiological methods, the wide array of inclusion criteria employed in the different studies, or previous administration of antibiotic therapy [36]. Pediatric patients with culture‐negative disease consistently show a trend towards younger age, lower body temperature, WBC count, and C‐reactive protein (CRP) values on admission, a milder clinical course, a shorter hospital stay, and a better prognosis, suggesting that fastidious pathogens of low virulence could be responsible for many of these infections [35,37–39]. The routine inclusion of sensitive NAATs in the workup of young children with suspected joint infections, and particularly those targeting specific K. kingae DNA sequences, has significantly improved the microbiological diagnosis of septic arthritis and confirmed that many culture‐negative cases of septic arthritis are caused by this microorganism [20,21]. However, even when sensitive NAATs are used, still one‐fifth of cases remain bacteriologically unconfirmed, indicating that many joint infections are caused by pathogens that are not detected by the current laboratory methods  [21]. Recently, metagenomic approaches based on shotgun next‐generation sequencing have been employed to identify the etiologic agents of culture‐negative prosthetic joint infections with good results [40]. The strategy can detect all the microorganisms that are present in a clinical specimen, including pathogens for which target sequences are not currently available. Despite these theoretical advantages, many technical problems remain to be solved, including the environmental background DNA contamination, the relative paucity of pathogen’s DNA in the synovial fluid specimen relative to the abundance of host’s sequences, need for extensive databases, and high cost [40]. It is to be expected that this and other future culture‐independent diagnostic methods will reduce or even eliminate the bacteriologically unconfirmed cases.

Pathogenesis The synovial membrane is highly vascular and lacks a limiting basement membrane, enabling easy bacterial access to the joint space in the course of a bacteremic episode. Once organisms have penetrated the joint, the low fluid shear conditions facilitate microbial adherence  [41]. Uncommonly, pediatric septic arthritis may also result from direct inoculation of organisms in the joint by a human or animal bite, joint taps especially with the injection of corticosteroids, or surgical procedures.

8  Native Joint Arthritis in Children

123

Invasion of the joint space in neonates occurs in the majority of cases as the result of the dissemination of the infection from a contiguous metaphyseal focus of osteomyelitis  [42]. In young children, the cartilaginous epiphyses receive their blood supply from a metaphyseal capillary network that obliterates between 6 and 9 months of age. Therefore, infection of a metaphyseal site can easily spread across the growth plate to the epiphysis and joint space. Because in older children the epiphyses and metaphyses have a separate blood supply and only the metaphyses of the hip, shoulder, and ankle bones remain intracapsular, the spread of infection from bone to joints becomes less common [42]. Occasionally, the neonatal joint may be directly invaded during a bacteremic episode and, in the hospital setting, by direct inoculation of skin organisms during a femoral venipuncture [43]. The source of the preceding bacteremia may be the result of nosocomial transmission of virulent S. aureus, the newborn’s normal skin flora, or acquisition of maternal organisms when delivered through a birth canal colonized by Enterobacteriaceae, S. agalactiae, or N. gonorrhoeae [4,12,24]. Bacteria implicated in septic arthritis usually display a variety of surface‐exposed receptors that recognize adhesive matrix molecules, such as collagen, fibronectin, and elastin, facilitating invasion by firmly anchoring the organism to the synovial layer [38]. Local trauma may unveil these tissue components, promoting bacterial adherence, and increasing the risk of suppurative arthritis. Inactivation of the genes encoding bacterial adhesins significantly reduces the capability of the organisms to establish a joint infection [41]. Both bacterial factors and the host’s immune response contribute to the progressive destruction of joint tissues [41]. Bacteria such as S. aureus are internalized by osteoblasts, causing apoptosis or evading the immune response by surviving and multiplying in the intracellular milieu and releasing potent toxins and enzymes that break down host tissues and provide nutrients for bacterial growth. The presence of bacteria in the joint induces a strong inflammatory response consisting of proliferation of the synovial cells, leukocyte migration, and formation of granulation tissue and abscesses. Synoviocytes and infiltrating leukocytes release proteases and secrete cytokines such as interleukin‐1‐β, interleukin‐6, and tumor necrosis factor‐α [41]. These cytokines activate an inflammatory cascade releasing acute‐phase reactants from the liver, such as CRP, that adhere to invading bacteria and facilitate opsonization and complement activation. On the other hand, cytokines increase the release of host matrix metalloproteinases, such as stromelysin, and other collagen‐degrading enzymes. The inflammatory process triggers fluid accumulation, increasing intra‐articular pressure and inducing tissue ischemia and necrosis [41]. The resulting cartilage destruction causes narrowing of the joint space and further erosive damage, leading to disabling orthopedic sequelae.

Clinical Presentation Typically, septic arthritis exhibits a more acute presentation than osteomyelitis, and most children with joint infections are brought to medical attention within two to five days from the onset of symptoms. Hematogenous pediatric septic arthritis affects a single joint in 95% of cases. The involvement of multiple articulations suggests a viral, reactive arthropathy or an immunocompromising condition. Polyarticular septic arthritis, however, has been noted in neonates, in half of the cases caused by gonococci, in 7% of those caused by S. aureus or H. influenzae type b, and in infections by Candida species [4].

124

Bone and Joint Infections

Septic arthritis usually affects the large weight‐bearing joints of the lower extremities (Figure 8.1). Small joints of the hand and feet are overrepresented in K. kingae infections, the sacroiliac joints are typically affected in brucellosis, and the sternoclavicular articulation by P. aeruginosa in intravenous‐drug users [26] and as a rare complication of subclavian vein catheterization [44] (see Chapter 10). Most children with septic arthritis present with acute onset of fever and local inflammatory changes, such as swelling or localized erythema of the overlying skin. Irritability, pain, abnormal (antalgic) posture, restricted range of motion or refusal to move the affected extremity or bear weight, and limping are frequent complaints. The pain of untreated septic arthritis is continuous and progressive, in contrast to inflammatory arthropathies such as juvenile idiopathic arthritis, where symptoms worsen upon rising in the morning. Infected joints are splinted by muscle contraction to limit motion and reduce pressure and the resulting pain. When the hip joint is involved, the extremity is held in flexion, external rotation, and abduction, the infected knee or ankle in slight flexion, and the shoulder in adduction and internal rotation. While examining the child, it should be kept in mind that arthritis of the hip is frequently difficult to localize and patients may present with pain referred to the knee or anterior thigh  [4]. Patients with sacroiliitis exhibit a positive FABERE (flexion, abduction, external rotation, extension) test. Painful palpation of the joint may be also elicited by direct compression of the iliac wing or by digital dorsal compression in rectal examination. Newborns and young patients infected with

Sternoclavicular 0.1%

Acromioclavicular 0.1% Metacarpal 0.1%

Shoulder 4.7% Elbow 14.0% Wrist 4.4% Interphalangeal 0.5%

Sacroiliac 0.6% Hip 22.2%

Knee 39.6% Metatarsal 0.4% Ankle 13.3%

Figure 8.1.  Anatomical distribution of 781 septic joints diagnosed in 725 children.

8  Native Joint Arthritis in Children

125

low‐grade virulence pathogens such as K. kingae or Brucella spp. may be afebrile at the time of diagnosis, requiring an increased awareness of the possibility of a joint infection [20,27]. In neonates, and especially in premature babies, the clinical picture may be dominated by nonspecific signs such as poor feeding, vomiting, abdominal distention, tachycardia, tachypnea, hypothermia, irritability or apathy, hypotension, poor perfusion, and acidosis  [33]. Meticulous physical examination may disclose limited use of an extremity or pseudoparalysis, and subtle signs of local inflammation over the affected joint, such as discomfort when handled or having the diaper changed, or swelling of the buttock, genitalia, thigh, or the entire extremity. In addition to obtaining synovial fluid specimens for culture, a complete sepsis workout, including obtaining blood and urine cultures and performance of a lumbar puncture, are indicated before administering empiric broad‐spectrum antimicrobial therapy.

Laboratory Investigation The key to the diagnosis of bacterial arthritis in children is a high index of clinical suspicion. The diagnosis should be confirmed without delay by aspiration of the joint, performed with a large‐bore needle (20‐gauge or larger). Although a comprehensive microbiological, biochemical, and cytological study of the synovial fluid is usually ordered [45], the only definitive proofs of an infectious etiology of the joint inflammation are either demonstration of bacteria in the Gram’s‐stain, growth of an unambiguous pathogen in culture, or detection of pathogen‐specific DNA sequences by a NAAT. If synovial fluid cannot be obtained by close needle aspiration, the procedure should be attempted again with imaging guidance, especially for sites that are not easily accessible such as the hips, shoulders, or sacroiliac joints [45]. The current microbiological approach for diagnosing pediatric septic arthritis is summarized in Figure  8.2. This laboratory strategy integrates the traditional Gram staining and culture on solid media, as well as novel and improved detection methods such as inoculation of the synovial fluid aspirate on blood culture vials and the use of sensitive NAATs. The algorithm takes into consideration the age of the child, and the presence of specific risk factors such as underlying immune deficiency, exposure to zoonotic pathogens, human or animal bites, invasive orthopedic procedures, etc. Aspiration of an amount of fluid insufficient for an extensive laboratory workup is common in young children or when a small joint is drained. In this case, the performance of a Gram’s‐stain, real‐time PCR assays with “universal” bacterial primers and K. kingae‐ specific primers in children aged 6–48 months, and inoculation of a blood culture vial are the best options. Any cloudy joint effusion should be considered infectious until proven otherwise. Although acute rheumatic fever, Reiter’s disease, and juvenile idiopathic arthritis can cause a markedly inflammatory synovial fluid, the highest leukocyte counts are seen in patients with septic arthritis, usually in the 50,000 to 200,000 cells per mm3 range, of which more than 90% are polymorphonuclear leukocytes. A WBC count higher than 50,000 leukocytes per mm3 is generally proposed as a cutoff to differentiate septic arthritis from non‐infectious joint exudates. Yet lower counts may be seen in infections caused by Gram‐negative organisms such as N. gonorrhoeae, K. kingae, and Brucella species, early in the course of bacterial arthritis of any etiology, and in neutropenic patients [46,47]. Conversely, WBC counts >50,000/mm3 of synovial fluid may be observed in children

126

Bone and Joint Infections

suspected septic arthritis in a child

specific risk factors*?

Predisposing condition

age group

yes

no

6 months–4 years

4 years Gram-stain

specimen

oropharyngeal

lab test

K. kingae-specific NAAT

+

Gram-stain

skeletal exudate

blood sample

culture on BCV

conventional culture

skeletal exudate

16S rRNA gene NAAT

– *: immunodeficiency, hemoglobinopathies, exposure to zoonotic infections, bites, puncture wounds, etc.

Figure 8.2.  Microbiological approach to the diagnosis of pediatric septic arthritis. NAAT: nucleic acid amplification test; BCV: blood culture vial.

with juvenile idiopathic arthritis, serum sickness, or reactive arthritis. Measurements of the synovial fluid glucose, protein, or lactate contents are neither sensitive nor specific for bacterial arthritis [48]. The aspirate should be transported to the microbiology laboratory without delay in the original syringe or a sterile tube. The use of swabs, although inexpensive and easy to use, should be discouraged. They are more likely to be contaminated; certain fibers, such as cotton, may inhibit bacterial growth; and organisms may remain adherent to swabs resulting in a false‐negative Gram’s‐stain examination and reducing the culture’s sensitivity [46]. A Gram’s stain should be prepared from a centrifuged synovial specimen and carefully examined. The test is positive in 75% of patients with staphylococcal arthritis but in less than half of those infected by Gram‐negative organisms  [45], probably because of a lower bacterial load and the difficulties in recognizing the presence of bacteria against the pink‐stained fibrin and cell background. The fluid should be seeded onto appropriate media (including a chocolate‐agar plate), and incubated in a CO2‐enriched atmosphere to enable the growth of capnophilic bacteria, such as pneumococci or neisseriae. Inoculation of a pediatric blood culture vial, preferably one containing antibiotic‐binding resins such as the BACTEC 9240 Peds Plus bottle  [49] or the BacT/Alert vial  [50], is also recommended because this method significantly improves the recovery of fastidious organisms such as K. kingae [20], and for patients receiving antimicrobial therapy. Routine cultures for anaerobic organisms, acid‐ fast bacilli, and fungi are not routinely indicated in children unless there are associated risk factors such as penetrating wounds, bites, or immunodeficiency [51].

8  Native Joint Arthritis in Children

127

Nowadays, NAATs are making a profound impact on the management of septic arthritis. This novel technology improves the detection of difficult‐to‐culture organisms, allows diagnosis in patients already treated with antibiotics, reduces time‐to‐detection, and enables precise identification of unusual pathogens  [52]. Although most clinical microbiology laboratories employ amplification of the universal 16S rDNA gene followed by sequencing of the amplicon, species‐specific primers that detect the most plausible pathogens have improved sensitivity and are less prone to contamination, but require epidemiological and clinical expertise to target the most probable bacterial targets (see Chapter 4). Since K. kingae is the prime agent of suppurative arthritis in children aged 6–48 months, and pediatric small joints and those that are not easily accessible are rarely taped, the performance of a sensitive K. kingae‐specific NAAT on an oropharyngeal specimen has been recommended for establishing the diagnosis in young patients [53]. The procedure is simple and non‐invasive, results are obtained in a few hours, and it could replace surgical diagnostic procedures. However, because a substantial fraction of young children carries the organism, a positive test result is not irrefutable proof of the etiology of the infection. However, the negative predictive value of the assay is very high and a negative PCR result practically excludes K. kingae as the possible etiology [53]. Blood cultures should be obtained in all children with suspected suppurative arthritis, not only because of the convenient accessibility of the specimen compared to obtaining a joint aspirate, but also because the etiologic agent may be recovered from the bloodstream in up to 50% of cases, even when cultures of synovial exudates are sterile [45]. When a gonococcal infection is suspected in an adolescent, cultures from the cervix, urethra, rectum, and oropharynx, as well as sensitive nucleic acid amplification tests of the urine, should also be performed [24]. The use and limitations of measuring levels of WBC and acute phase reactants such as erythrocyte sedimentation rate, C‐reactive protein, and procalcitonin levels for diagnosing septic arthritis are discussed in Chapter 9.

Imaging Studies Plain radiographs have a low sensitivity for detecting joint effusions, and thus are not diagnostic for septic arthritis help, but may help to detect concomitant osteomyelitis and exclude other conditions. Ultrasound has the advantage of being a non‐invasive technique, and is especially helpful for deep joints such as the hip, although its performance is operator‐dependent. Ultrasound may detect the accumulation of intra‐articular fluid, and guide the performance of a diagnostic joint aspiration. Neither the size nor the echogenicity, however, allows to firmly conclude whether the detected effusion is infected or not. Conversely, the absence of joint fluid accumulation in the hip joint space helps to exclude septic arthritis. Bone scans may be used to localize the joint affected when in doubt and detect unsuspected multifocal disease. The characteristic finding in septic arthritis consists of increased uptake on both sides of the joint during the early phase, whereas in osteomyelitis, unilateral increased uptake is observed [54]. It should be pointed out that the interpretation of bone scans is difficult in neonates. Magnetic resonance imaging (MRI) provides high contrast and resolution and multiplanar imaging. These technical features result in a superb definition of the extent

128

Bone and Joint Infections

of soft tissue and cartilage involvement. Routine use of the MRI frequently reveals the concomitant involvement of the bone and the adjacent joint [55]. The technique is useful in differentiating suppurative arthritis from transient, non‐infectious synovitis of the hip. Demonstration of contralateral (asymptomatic) joint effusion and signal intensity alterations and enhancement in surrounding soft tissues are associated with transient synovitis  [56], while edema of the adjacent bone is exclusively associated with septic arthritis [57]. The MRI, however, has the limitations of costs and availability, and the need to sedate the young and uncooperative child [58].

Differential Diagnosis The differential diagnosis of septic arthritis is wide and depends on the patient’s age, underlying clinical conditions, and the joint involved (Table 8.3).

Treatment Because of the potential risk for long‐term disabilities caused by the rapid destruction of the joint cartilage, septic arthritis should be considered a true emergency. Optimal management of suppurative arthritis of childhood requires a combination of medical and surgical interventions that is accomplished best through the joint efforts of pediatricians and experienced orthopedic surgeons. A high index of clinical suspicion, drawing blood and synovial fluid samples for bacteriological diagnosis, prompt joint drainage, and administration of adequate antimicrobial therapy are the cornerstones of an optimal treatment. Table 8.3.  Differential diagnosis of pediatric septic arthritis. • Osteomyelitis • Viral arthritis (parvovirus B19, HTLV‐1, HIV, rubella, alphaviruses, hepatitis B and C) • Reactive arthritis • Transient synovitis • Rheumatic fever • Bone infarction • Autoimmune arthritis • Septic bursitis • Sickle cell anemia • Slipped capital femoral epiphysis • Perthes disease • Villonodular synovitis • Spondylodiscitis • Psoas abscess • Serum sickness • Henoch‐Schönlein purpura • Hemophilia • Trauma • Familial Mediterranean fever • Chronic recurrent multifocal osteomyelitis • Leukemia • Neuroblastoma

8  Native Joint Arthritis in Children

129

Joint Space Drainage Evacuation of the joint space by either close needle aspiration, arthroscopy, or surgical drainage provides synovial fluid samples for diagnostic purposes, reduces intracapsular pressure relieving pain, and removes bacteria and cartilage‐damaging toxin products [59]. There is still controversy over the best mode of drainage, but most researchers have concluded that repeated aspirations of joints other than the hip is associated with a superior outcome as compared to arthrotomy  [45,60,61]. However, because of the retrospective nature of most studies, a selection bias, in which more clinically severe cases could have been selected for open surgery and, naturally, experienced a worst outcome, cannot be excluded. The customary indications for surgical drainage are summarized in Table 8.4. Because the femoral head receives its blood supply from a single intra‐articular arterial branch, accumulation of pus within the hip joint may result in increased pressure and vascular occlusion, compromising bone tissue viability. Avascular necrosis of the femoral head, joint instability, premature epiphyseal closure, and limb‐length discrepancy are common complications of delayed or inadequate treatment of hip joint infections, necessitating complex corrective surgery, usually with dismal results  [62]. Therefore, prompt open surgical drainage has been traditionally recommended [63]. In recent years, performance of ultrasound‐guided aspiration of the joint space has been advocated instead  [64]. After the aspiration, the joint is irrigated using the same needle and the procedure is repeated daily for three to five days. Only 4 of 28 patients treated with this modality did not improve and required surgical drainage, and no complications were detected after a mean follow‐up of over seven years  [62]. In a second study, three‐ directional arthroscopic drainage and lavage were performed in children older than 6 years with staphylococcal hip arthritis with excellent functional outcomes [65]. Despite these encouraging results, additional experience with this unorthodox treatment is needed before surgery‐sparing approaches can be routinely recommended for pediatric suppurative arthritis of the hip.

Antibiotic Therapy The choice of the initial antibiotic therapy (pending culture and/or nucleic‐acid amplification assays results) should be guided by the Gram’s‐stain examination of the fluid, patient’s age, vaccination status, presence of specific risk factors (such as immunodeficiency), potential exposure to organisms such as B. burgdorferi or Brucella spp., and the local prevalence of antibiotic resistance in relevant organisms such as S. aureus [66]. Guidelines for the initial antibiotic administration are provided in Table 8.5. Large doses of β‐lactamase‐resistant β‐lactam drugs, cephalosporins, or clindamycin, should be used by either the oral or the parenteral route.

Table 8.4.  Indications for arthrotomy in pediatric septic arthritis. • Arthritis of the hip and shoulder • Presence of large amounts of fibrin, debris, or loculation within the joint space • Presence of an implant • Arthritis not responding to medical treatment within three days

130

Bone and Joint Infections

Table 8.5.  Guidelines for administration of initial antibiotic therapy, pending culture results [64]. Dosage Age group

Antibiotic

mg/kg/day

doses/day

Neonate

BLRBL or Vancomycin or Clindamycin plus Cefotaxime

100   30   20–30   100–150

4   2   3   3

Child ≤4 years

BLRBL or Vancomycin or Clindamycin plus 1st or 2nd generation cephalosporin or Ceftriaxone

150   40–60   40   150   100

4   3   4   4   1

Child >4 years

BLRBL or Vancomycin or Clindamycin

150   45–60   30–40

4   3   3

Sexually active adolescent

BLRBL or Vancomycin or Clindamycin plus Ceftriaxone

150   40–60   30–40   100

4   3   3   1

BLRBL: β‐lactamase‐resistant β‐lactam (nafcillin, cloxacillin, flucloxacillin, or dicloxicillin).

Ceftriaxone offers the advantages of a wide antimicrobial spectrum, once‐a‐day dosage, good safety profile, and comparable results to those obtained with oxacillin in a retrospective study of adult skeletal infections [67]. However, because of its high serum protein binding (>90%) and relative high MIC compared to that of β‐lactamase‐resistant penicillins such as oxacillin or first generation cephalosporins, there is reluctance to use it alone for suspected or culture‐proven methicillin‐susceptible S. aureus arthritis in children. In patients under 4 years of age, a combination of an antistaphylococcal penicillin and a broad‐spectrum cephalosporin (cefotaxime in the newborn, and ceftriaxone or cefuroxime in older children) will provide adequate initial therapy for the most common bacterial pathogens. In premature babies and in those receiving intensive care, empiric therapy against nosocomial bacteria and yeasts should be considered. Over the age of 4 years, coverage against Gram‐positive bacteria with narrow‐spectrum antibiotics, such as a β‐ lactamase‐resistant penicillin, vancomycin, or clindamycin, is usually administered, unless an immunocompromising condition is present. It should be pointed out that in

8  Native Joint Arthritis in Children

131

areas where CA‐ MRSA is prevalent (>10% of staphylococcal isolates), vancomycin or clindamycin should be used instead of the β‐lactamase‐resistant penicillin, pending culture results (Figure 8.3). Co‐trimoxazole (16/90 mg/kg/day, b.i.d.) is an alternative that is active against most CA‐MRSA, provides adequate wide antimicrobial coverage, and has an excellent oral bioavailability [66]. Because of a scarcity of well‐designed, randomized, controlled, and sufficiently powered studies, no evidence‐based recommendations on the duration of antibiotic therapy for pediatric septic arthritis can be formulated. Most accepted treatment protocols are based on retrospective case analyses and personal experience and, traditionally, prolonged in‐hospital administration of intravenous antibiotics has been advocated. Usually, the total duration of the prescribed therapeutic courses varied between three to six weeks, depending on the patient’s age, bacterial identity (longer for S. aureus and enterobacteriaceae), and location of the infected site. This orthodox concept has been slowly evolving toward shorter antibiotic regimens, sequential parenteral‐oral therapy, and early hospital discharge, if the following criteria are met: i) the child is able to take oral medications; ii) the identity of the causative organism is known; iii) an oral agent with excellent bioavailability is available; iv) the patient’s compliance can be ascertained; v) the family can be relied on to adhere to the antibiotic regimen at home; and vi) levels of CRP decrease and can be monitored  [67–69]. The advantages of this approach are saving of hospitalization days, reduced treatment costs, lesser disruption of family life, shortened exposure to health care‐associated infections, and avoidance of the untoward effects of prolonged parenteral antibiotic therapy [70]. Although in the past, peak serum bactericidal titers of the oral antibiotic ≥1:8 were required to switch to oral therapy [70], suspected septic arthritis

MRSA >10% of Staphylococcus aureus?

no

yes

clindamycin resistance in MRSA: 50,000 cells/μl), and/or a predominance of granulocytes (>90%) in synovia in the absence of crystal disease. Eubacterial polymerase chain reaction (PCR) yields a lower sensitivity than standard culture, and is still relatively expensive. In polymicrobial infections, its interpretation may be difficult. Moreover, it does not provide information about antibiotic resistance, except for genes coding for methicillin or rifampin resistance. In contrast, specific or multiplex PCR is beneficial in special circumstances in which slow‐growing bacteria are suspected, such as Kingella kingae  [18], Brucella spp., Coxiella burnetii, Bartonella henselae, M. tuberculosis, or M. ulcerans. In case of suspicion of gonococcal disease, Neisseria gonorrhoeae should be searched with a specific PCR in the urine and/or in synovial fluid [15,16].

9  Native Joint Arthritis in Adults

143

Inflammatory Parameters in Serum In localized arthritis without concomitant bacteremia, procalcitonin is frequently negative despite the presence of localized, purulent, bacterial infection [19–21]. Serum procalcitonin is very likely positive in concomitant sepsis and/or bacteremia  [22]. However, practically all serum inflammatory markers are elevated in systemic infections. This fact has implications. As many primary bacterial arthritis might have an hematogenous origin, it is not surprising that these serum inflammatory markers probably reflect rather the silent concomitant bacteremia than the mere arthritis itself [2]. When comparing the serum inflammatory markers among each other, serum procalcitonin is a useful marker with diagnostic performances higher than those of other biomarkers (white blood cells, C‐reactive protein) [20,21].

Synovial Fluid Analysis Inflammatory markers can also be tested locally, i.e. in synovial fluid. There has been a panoply of publications investigating their different thresholds to assess the bacterial nature of infection. For example, Baillet et al. [23] determined whether calprotectin and a‐defensins could discriminate septic from other inflammatory arthritis. Among 73 patients, calprotectin was the only biomarker that discriminated septic arthritis from non‐septic inflammatory arthritis, with 76% sensitivity, 94% specificity, and a positive likelihood ratio of 12.2 at the threshold for calprotectin of 150 mg/l [23]. In the future, more intrasynovial markers may emerge. As an example, serum IL‐6 levels are sensitive and specific in the diagnostic evaluation of periprosthetic joint infection [24]. There are no data on its use in patients with native arthritis. However, it can be predicted that it would not be discriminative against autoimmune inflammatory arthritis [25]. Regarding the threshold of synovial leukocytes, many experts consider a number ≥50,000 cells/μl as predictive for a septic origin [26]. Others indicate that even 100,000 cells/μl fail to discriminate septic from non‐septic joint infections [27]. A literature review, including 6242 patients, indicates that a neutrophil proportion >90% reveals a threefold likelihood for bacterial origin of arthritis  [28]. The literature cannot provide an unambiguous threshold. In our opinion, no intra‐articular cell counts or neutrophil fraction are accurate enough to determine a therapeutic approach basing on these laboratory results only (Figure 9.1).

Bacterial Staining on Microscopy Even very few bacterial counts in the otherwise sterile synovial fluid are sufficient to trigger a strong local and systemic inflammatory response. However, this low bacterial inoculum classically remains undetected in microscopic examination  [29]. While many centers still use Gram‐ or acridine orange‐staining for a rapid diagnosis of bacterial arthritis, the cost–benefit ratio of such an approach is very probably negative. In a large single‐center study, we assessed 500 different bacterial arthritis episodes. The Gram‐ staining revealed pathogens in a total of 146 cases (146/500, 29%), or in only 146/400 (37%) microbiologically confirmed culture‐positive cases. Overall, the sensitivity, specificity, and positive and negative predictive values of Gram‐stain for the diagnosis of septic arthritis were 0.37, 0.99, 0.99, and 0.28, respectively, when culture positivity was defined as gold standard. Quite similar values were recorded across different patient subpopulations, in particular for sensitivity values that were 0.40 for immune‐compromised

144

Bone and Joint Infections

patients, 0.36 for patients who were treated with antibiotics, and 0.52 for those with ­concomitant crystal arthropathy. A CRP level above 150 mg/l, the presence of bacteremia, and a synovial leukocyte count >180,000/μl were significantly linked with “positive Gram‐staining.” In contrast, the predominance of intrasynovial neutrophils was not ­predictive for a positive Gram‐staining [29].

Imaging In contrast to chronic osteomyelitis (Chapter 20), imaging procedures are of low diagnostic value for native bacterial arthritis, even if radiographs and ultrasound may reveal ­concomitant soft tissue swelling or hidden abscesses outside the infected joint. However, imaging reveals the septic nature of arthritis only in case of concomitant osteomyelitis. Ultrasound may be useful to facilitate arthrocentesis, especially in hip arthritis.

Treatment General Aspects The therapeutic management of native joint bacterial arthritis in adult patients is mostly based on experts’ opinion and individualized local experience. It is not standardized because of variable clinical presentations and the lack of sufficient scientific data derived from prospective‐randomized trials. Therefore, unsurprisingly, treatment differs in many parts of the world and between academic settings. Therapeutic analogies to the management of other bone and implant‐related infections are common. As a general consensus, the therapy relies on two principles, namely lavage/drainage and systemic antibiotics [2]. In contrast, the degree of emergency of the joint drainage is still a matter of discussion. Many surgeons and physicians emphasize that every bacterial arthritis, regardless of its origin and clinical presentation, should be drained as soon as possible. This opinion is based on animal studies, showing that very early antimicrobial therapy is needed to prevent cartilage destruction in staphylococcal arthritis of rabbits [30]. Despite these experimental data, some physicians favor less urgent therapy, in order to wait for optimal conditions for drainage. In a patient without clinical sepsis, they propose to start empirical antimicrobial therapy, but to postpone drainage, until an orthopedic surgeon or rheumatologist is available. Thus, they prefer not performing drainage as an emergency procedure, e.g. during night shifts. Lauper et al. [31] analyzed over 200 native adult septic arthritis cases, involving interdigital hand and foot joints (n=46), knees (n=67), shoulders n=48), hips (n=22), and ankles (n=8). Patients with lavage within 6 hours of admission had similar long‐term functional outcomes as compared to those with drainage between 6–12h, 12–24h, or even after more than 24h. Interestingly, the median duration of symptoms before hospital admission was 3 days, meaning that an early emergency surgery only shortened the total duration of disease by at maximum 10–25%  [31]. Indeed, in available literature and real‐life conditions, deleterious delays in adult patients are longer, and do not rely on the time in the emergency room. For example, according to Vispo‐ Seara et al. [32], an advanced cartilage damage significantly correlates with a cutoff of >2 weeks between symptoms and drainage. Ross et al. [33] failed to demonstrate that a short delay between admission and intervention influenced any outcome. For Balabaud et al. [34], the pre‐surgical delay was shorter (mean 12 days) in cured knee septic arthritis episodes compared to clinical failures (23 days). However, the authors equally failed to

9  Native Joint Arthritis in Adults

145

define a specific threshold. Klinger et al. [35] reported better scores if the local symptoms begun ≤2 weeks prior to shoulder drainage. Matthews et  al.  [36] identified a ≥3‐weeks delay as predictive for femoral head excision due to secondary osteitis.

Surgical Drainage versus Arthrocentesis While surgeons naturally prefer operative joint lavage [37], many rheumatologists (e.g. in France) perform serial closed needle aspirations of the joint for lavage in stable patients. A large, multicentric head‐to‐head comparison between surgery and repeated arthrocentesis is difficult because of the large case‐mix of adult arthritis patients and cultural differences in the management. Harada et al. [38] retrospectively compared the outcomes of 20 patients treated with bedside closed‐needle joint aspiration versus the one of 41 patients undergoing surgical (arthrotomy or arthroscopy) approaches. They evaluated time to recovery, disposition to rehabilitation, recurrence of infection, and mortality. There were no differences regarding all outcomes between both groups at twelve months’ follow‐up. However, patients managed by arthrocentesis experienced more frequent recovery at three months and needed less often short‐term rehabilitation. These results may be biased, since patients were not randomized to conservative versus surgical management. Thus, there is a high risk that patients with more severe arthritis were treated surgically. In addition, because none of the authors were surgeons, patients were obviously not managed with a multidisciplinary team. Similarly, Ravindran et al. [39] retrospectively compared 19 patients managed surgically (arthrotomy/arthroscopy) with 32 patients treated medically (serial closed‐needle aspiration) for native septic arthritis. Again, complete recovery did not differ. However, surgically drained patients also needed more physiotherapy for re‐education. Again, there is a risk for a bias in the analysis of this non‐randomized retrospective study. Taken together, the debate concerning the optimal drainage regarding technique (arthrocentesis, arthroscopy, arthrotomy) and number of drainages is not closed. We advocate against a planned second drainage (second look) solely basing on visual aspects during the first drainage; and propose to evaluate the need for a second intervention according to the clinical progress. A successful intervention usually should lead to pain reduction, reduced swelling, and less fever after two days following a drainage and with a correct antibiotic therapy.

Antibiotic Therapy Choice of Antimicrobial Agent Little is known about the penetration of antimicrobial agents into synovia. According to a recent review [40], studies on joint space penetration are fewer than studies on bone tissue penetration. The authors evaluated more than 30 antibiotics regarding bone and joint penetration. Overall, most antibiotics, including amoxicillin, piperacillin/tazobactam, aztreonam, cephalosporins, carbapenems, clindamycin, aminoglycosides, fluoroquinolones, doxycycline, vancomycin, linezolid, daptomycin, fosfomycin, co‐trimoxazole, rifampin, and dalbavancin, showed good penetration into osteoarticular tissues reaching concentrations exceeding the minimal inhibitory breakpoints of common pathogens. Few exceptions include penicillin and metronidazole, which showed a lower than optimum penetration into bones, and the latter as well as flucloxacillin had poor profiles in terms of joint space penetration. Although clinical studies in osteomyelitis and septic arthritis are not available for all of the evaluated antibiotics, pharmacokinetic results indicate that

146

Bone and Joint Infections

Table 9.2.  Antibiotic therapy for bacterial native joint infections (proposition of the authors). Parenteral therapy Microorganisms

Treatment of choice

Alternatives

Oral therapy

Methicillin‐susceptible S. aureus

Flucloxacillin 2g every 6h

Cefazolin 2 g every 8h or Cefuroxime 1.5 g every 8h

Clindamycin 600 mg every 8h

Methicillin‐resistant S. aureus

Vancomycin (15mg/ kg/12h)

Daptomycin 8 mg/ kg/24h

Co‐trimoxazole 1 ds tablet/12h

Streptococci

Penicillin G IV (3 million U 4–6x/d)

Ceftriaxone 2 g every 24 h

Amoxicillin 1 g/8h, Clindamycin 600mg/8h

Enteric Gram‐negative rods

Ceftriaxone 2 g every 24h

Cefuroxime 1.5 g IV every 8h

Ciprofloxacin 750 mg/12h

Serratia, P. aeruginosa

Ceftazidime 2 g every 8h

Cefepime 2 g every 12h

Ciprofloxacin 750 mg/12h

Anaerobes

Clindamycin 600 mg/8h

Empirical therapy

Clindamycin 600 mg/8h

Amoxicillin‐ clavulanic acid 2.2 g every 6–8h

Imipenem 500 mg every 6h

Cephalosporins 2nd/3rd generation

Amoxicillin‐ clavulanic acid 2.2 g every 6–8h

Metronidazole 500 mg/8h

agents with good penetration profiles would have a potential utilization in such ­infections  [40]. Table  9.2 shows the standard and alternative antimicrobial regimens against the most commonly encountered microorganisms. Importantly and preferably, β‐ lactam antibiotics should be avoided for the oral route (at least initially), because of their low oral bioavailability. Against methicillin‐resistant pathogens, vancomycin is the standard intravenous drug. While it is believed that trough serum levels of 20 mg/ml are required for optimal treatment of bone infections [41], there is no similar consensus for native joint bacterial arthritis. Vancomycin should be administered over at minimum one hour to ­prevent the histamine‐mediated “red man” syndrome. Fluoroquinolones are important agents against Gram‐negative infections. Because of their excellent bioavailability, they can be used by the oral route practically from the start. A multicenter Swiss study proved that oral combination therapy with fluoroquinolones plus rifampicin was an alternative to ­parenteral therapy (remission rates; 86% vs. 84%, respectively) for all staphylococcal infections, including arthritis in 35 cases [42]. Duration of Antimicrobial Therapy The optimal duration of post‐drainage systemic antibiotic treatment remains controversial, and might also, theoretically, depend on the individual case. Different regimens are (or have been) recommended such as three‐week IV therapy for streptococci, a total of IV

9  Native Joint Arthritis in Adults

147

for three to four weeks for staphylococci and Gram‐negatives, and more than four weeks IV for immune‐compromised patients  [2]. Most authors generally recommend an IV therapy for the initial two weeks, followed by an oral relay during two additional weeks  [43]. However, most experts consider this duration excessive for drained, adult native joint infections. A total treatment duration of three to four weeks can be regarded as sufficient for drained intrasynovial infections without bone involvement [2]. Most controlled trials, evaluating patients with septic arthritis, have been performed in children. Many pediatric publications show a successful outcome with a total antibiotic duration of 10–20 days, including only a short period of parenteral antibiotics [44]. The duration of antimicrobial therapy in bacterial hand and wrist arthritis is traditionally short. Angly et al. [45] reviewed 31 operated adult finger arthritis episodes. No recurrence occurred after antibiotics administered for a median duration of 2 days IV, and 17 days orally. In another study of 101 hand arthritis cases, the transition from IV to oral therapy occurred 3–5 days after surgery. All outcomes were similar [46]. The literature favouring this short duration for small joint infections of the hand is emerging and has been recently published in a narrative review [47]. Whether a short course is also sufficient in patients with large joint arthritis is not yet clear. We recently published a prospective‐randomized trial evaluating the duration of therapy in 154 adult patients with native joint bacterial arthritis [48]. However, only one‐ third of the patients suffered from medium or large joint infection. There were no differences regarding clinical remissions, adverse events, or mechanical sequels in patients treated with two weeks compared to those with four weeks of systemic targeted antibiotics after surgical drainage. These results were valid for both the intention‐to‐treat and the per‐protocol analyses, the entire arthritis population, or when analyzing the substratum of the hand arthritis cases separately. Three out of 154 episodes recurred, only one of them in the two‐week arm. Similar to the pediatric series, the trial also confirmed that a short initial IV‐therapy for arthritis is sufficient, because the median duration of initial IV therapy was only two days [48]. Taken together, two‐week antibiotic therapy after drainage can be considered in patients with septic arthritis, especially of the hand and wrist. However, before a two‐week course of antibiotics can be generally recommended for patients with large joint arthritis, larger controlled trials should confirm the non‐ inferiority of a short course in this population.

Outcome A fatal course of primary bacterial native joint arthritis is very rare, and rather caused by concomitant sepsis or severe underlying primary infection, such as endocarditis. The proportion of infection remission may reach 90–95% for large joints  [2] and 97% for small joints  [48]. In epidemiological studies among adult patients, the therapeutic variables are rarely associated with treatment failures. Rather, the comorbidities such as chronic steroid administration and organ transplantation are the strongest predictors of recurrence [2]. A major problem is the (mechanical) sequels of infection. Indeed, long‐term mechanical sequels are up to five times more prevalent than recurrent arthritis or reinfections [2,48]. Prospective‐randomized trials in adult patients with septic arthritis, assessing all adverse events, reveal 20 and 35% of mechanical sequels, of which 15% require another surgical intervention [31,48]. Besides re‐education, sensory‐integrative therapy, and eventual

148

Bone and Joint Infections

c­ orrective surgery, we currently lack measures to efficiently prevent these sequels, or to diminish their severity. Of note, radiologic changes such as secondary osteoarthritis (arthrosis) may occur as a result of cartilage damage by infection. These changes are not always associated to clinical impairment [31,32].

Key Points ●● ●●

●●

●●

All septic arthritis cases require drainage, either once or more often sequentially. Most cases of primary septic arthritis are due to Staphylococcus aureus. Posttraumatic cases show a higher proportion of Gram‐negative and atypical pathogens. The recommended duration of antibiotic therapy is two weeks for hand and wrist arthritis and three to four weeks for medium and large joint arthritis. A short initial parenteral therapy can be followed by oral antibiotics with good bioavailability. Mechanical sequels are a major and cumbersome problem, occurring in up to 30% of cases.

Acknowledgments We would like to thank the Department of Orthopedic Surgery of Balgrist University Hospital for support.

References 1. Elsissy JG, Liu JN, Wilton PJ, et al. Bacterial septic arthritis of the adult native knee joint: a review. JBJS Rev. 2020;8(1):0059. 2. Uçkay I, Tovmirzaeva L, Garbino J, et al. Short parenteral antibiotic treatment for adult septic arthritis after successful drainage. Int J Infect Dis. 2013;17(3):199–205. 3. Uçkay I, Hirose CB, Assal M. Does intra‐articular injection of the ankle with corticosteroids increase the risk of subsequent periprosthetic joint infection (PJI) following total ankle arthroplasty (TAA)? If so, how long after a prior intra‐articular injection can TAA be safely performed? Foot Ankle Int. 2019;40(1):3–4. 4. Ross JJ, Ard KL, Carlile N. Septic arthritis and the opioid epidemic: 1465 cases of culture‐positive native joint septic arthritis from 1990–2018. Open Forum Infect Dis. 2020;7(3):089. 5. Geirsson ÁJ, Statkevicius S, Víkingsson A. Septic arthritis in Iceland 1990‐2002: Increasing incidence due to iatrogenic infections. Ann Rheum Dis. 2008;67(5):638–643. 6. Favero M, Schiavon F, Riato L, et al. Septic arthritis: a 12 years retrospective study in a rheumatological university clinic. Reumatismo. 2008;60(4):260–267. 7. Di Benedetto C, Hoffmeyer P, Lew I, Uçkay I. Post‐traumatic septic arthritis. Eur Musculoskel Rev 2012;7. 8. Eder L, Zisman D, Rozenbaum M, et al. Clinical features and aetiology of septic arthritis in northern Israel. Rheumatology 2005;44:1559–63. 9. Weston VC, Jones AC, Bradbury N, et al. Clinical features and outcome of septic arthritis in a single UK Health District 1982–1991. Ann Rheum Dis. 1999;58(4):214–219. 10. Kennedy N, Chambers ST, Nolan I, et al. Native joint septic arthritis: Epidemiology, clinical features, and microbiological causes in a New Zealand population. J Rheumatol. 2015;42(12): 2392–2397.

9  Native Joint Arthritis in Adults

149

11. Clerc O, Prod’hom G, Greub G, et al. Adult native septic arthritis: a review of 10 years of experience and lessons for empirical antibiotic therapy. J Antimicrob Chemother. 2011;66(5):1168–1173. 12. Kaandorp CJE, Dinant HJ, van de Laar MAFJ, et  al. Incidence and sources of native and prosthetic joint infection: A community based prospective survey. Ann Rheum Dis. 1997;56(8):470–475. 13. Stutz G, Kuster MS, Kleinstück F, et al. Arthroscopic management of septic arthritis: Stages of infection and results. Knee Surg Sports Traumatol Arthrosc. 2000;8(5):270–274. 14. Uçkay I, Sax H, Harbarth S, et al. Multi‐resistant infections in repatriated patients after natural disasters: lessons learned from the 2004 tsunami for hospital infection control. J Hosp Infect. 2008;68(1):1–8. 15. Rice PA. Gonococcal arthritis (disseminated gonococcal infection). Infect Dis Clin North Am. 2005;19(4):853–861. 16. García‐De La Torre I, Nava‐Zavala A. Gonococcal and nongonococcal arthritis. Rheum Dis Clin North Am. 2009;35(1):63–73. 17. Al‐Mayahi M, Cian A, Lipsky BA, et al. Administration of antibiotic agents before intraoperative sampling in orthopedic infections alters culture results. J Infect. 2015;71(5):518–525. 18. Samara E, Spyropoulou V, Tabard‐Fougère A, et al. Kingella kingae and osteoarticular infections. Pediatrics. 2019;144(6):20191509. 19. Uçkay I, Garzoni C, Ferry T, et al. Postoperative serum pro‐calcitonin and C‐reactive protein levels in patients with orthopedic infections. Swiss Med Wkly. 2010;140:13124. 20. Chouk M, Verhoeven F, Sondag M, et al. Value of serum procalcitonin for the diagnosis of bacterial septic arthritis in daily practice in rheumatology. Clin Rheumatol. 2019;38(8): 2265–2273. 21. Paosong S, Narongroeknawin P, Pakchotanon R, et al. Serum procalcitonin as a diagnostic aid in patients with acute bacterial septic arthritis. Int J Rheum Dis. 2015;18(3):352–359. 22. Jung SW, Kim DH, Shin SJ, et al. Septic arthritis associated with systemic sepsis. Int Orthop. 2018;42(1):1–7. 23. Baillet A, Trocmé C, Romand X, et al. Calprotectin discriminates septic arthritis from pseudogout and rheumatoid arthritis. Rheumatology (Oxford). 2019;58(9):1644–1648. 24. Gallo J, Svoboda M, Zapletalova J, et al. Serum IL‐6 in combination with synovial IL‐6/CRP shows excellent diagnostic power to detect hip and knee prosthetic joint infection. PLoS One. 2018;13(6):0199226. 25. Ogata A, Kato Y, Higa S, et al. IL‐6 Inhibitor for the treatment of rheumatoid arthritis: a comprehensive review. Mod Rheumatol. 2019;29(2):258–267. 26. McGillicuddy DC, Shah KH, Friedberg RP, et  al. How sensitive is the synovial fluid white blood cell count in diagnosing septic arthritis? Am J Emerg Med. 2007;25(7):749–752. 27. Siva C, Velazquez C, Mody A, et  al. Diagnosing acute monoarthritis in adults: A practical approach for the family physician. Am Fam Physician. 2003;68(1):83–90. 28. Margaretten ME, Kohlwes J, Moore D, et  al. Does this adult patient have septic arthritis? JAMA. 2007;297(13):1478–1488. 29. Cunningham G, Seghrouchni K, Ruffieux E, et al. Gram and acridine orange staining for diagnosis of septic arthritis in different patient populations. Int Orthop. 2014;38(6):1283–1290. 30. Smith RL, Schurman DJ, Kajiyama G, et al. The effect of antibiotics on the destruction of cartilage in experimental infectious arthritis. J Bone Joint Surg Am. 1987;69(7):1063–1068. 31. Lauper N, Davat M, Gjika E, et al. Native septic arthritis is not an immediate surgical emergency. J Infect. 2018;77(1):47–53. 32. Vispo Seara JL, Barthel T, Schmitz H, et al. Arthroscopic treatment of septic joints: Prognostic factors. Arch Orthop Trauma Surg. 2002;122(4):204–211. 33. Ross JJ, Saltzman CL, Carling P, et al. Pneumococcal septic arthritis: review of 190 cases. Clin Infect Dis. 2003;36(3):319–327. 34. Balabaud L, Gaudias J, Boeri C, et al. Results of treatment of septic knee arthritis: A retrospective series of 40 cases. Knee Surg Sports Traumatol Arthrosc. 2007;15(4):387–392.

150

Bone and Joint Infections

35. Klinger HM, Baums MH, Freche S, et al. Septic arthritis of the shoulder joint: An analysis of management and outcome. Acta Orthop Belg. 2010;76(5):598–603. 36. Matthews PC, Dean BJF, Medagoda K, et  al. Native hip joint septic arthritis in 20 adults: Delayed presentation beyond three weeks predicts need for excision arthroplasty. J Infect. 2008;57(3):185–190. 37. Wirtz DC, Marth M, Miltner O, et  al. Septic arthritis of the knee in adults: Treatment by arthroscopy or arthrotomy. Int Orthop. 2001;25(4):239–241. 38. Harada K, McConnell I, Derycke EC, et al. Native joint septic arthritis: Comparison of outcomes with medical and surgical management. South Med J. 2019;112(4):238–243. 39. Ravindran V, Logan I, Bourke BE. Medical vs surgical treatment for the native joint in septic arthritis: a 6‐year, single UK academic centre experience. Rheumatology (Oxford). 2009;48(10):1320–1322. 40. Thabit AK, Fatani DF, Bamakhrama MS, et al. Antibiotic penetration into bone and joints: an updated review. Int J Infect Dis. 2019;81:128–136. 41. Hidayat LK, Hsu DI, Quist R, et al. High‐dose vancomycin therapy for methicillin‐resistant Staphylococcus aureus infections: efficacy and toxicity. Arch Intern Med. 2006;166(19): 2138–2144. 42. Schrenzel J, Harbarth S, Schockmel G, et al. A randomized clinical trial to compare fleroxacin‐ rifampicin with fucloxacillin or vancomycin for the treatment of staphylococcal infection. Clin Infect Dis. 2004;39(9):1285–1292. 43. Cho HJ, Burke LA, Lee M. Septic arthritis. Hospital Medicine Clinics. 2014; 3:494–503. 44. Peltola H, Paakkonen M, Kallio P, et al. Prospective, randomized trials of 10 days versus 30 days of antimicrobial treatment, including a short‐term course of parenteral therapy, for childhood septic arthritis. Clin Infect Dis. 2009;48(9):1201–1210. 45. Angly B, Steiger R, Zimmerli W. Infektiöse Arthritis der Fingergelenke. Handchir Mikrochir Plast Chir. 2007;39(2):118–123. 46. Meier R, Wirth T, Hahn F, et al. Pyogenic arthritis of the fingers and the wrist: Can we shorten antimicrobial treatment duration? Open Forum Infect Dis. 2017;25;4:058. 47. Sendi P, Kaempfen A, Uçkay I, et al. Bone and joint infections of the hand. Clin Microbiol Infect. 2020;1198:848–856. 48. Gjika E, Beaulieu JY, Vakalopoulos K, et al. Two weeks versus four weeks of antibiotic therapy after surgical drainage for native joint bacterial arthritis: a prospective, randomised, non‐ inferiority trial. Ann Rheum Dis. 2019;78(8):1114–1121.

Chapter 10

Septic Arthritis of Axial Joints Werner Zimmerli

In this chapter, arthritis in three different axial joints will be presented, namely sternoclavicular arthritis, symphysitis pubis, and septic sacroiliitis. All three are cartilaginous joints with a small range of motion [1]. Septic arthritis of these three joints is rare, and generally, but not exclusively, occurs in patients with special risk factors. The sternoclavicular articulation is a gliding diarthrodial joint  [2,3]. Infection of this articulation can present as arthritis with a minimal amount of synovial empyema, as soft tissue abscess or as concomitant osteomyelitis. In contrast, the pubic symphysis and the sacroiliac joint are amphiarthrotic, which means only slightly movable. The former is a synchondrosis, the latter a syndesmosis. A synchondrosis is a cartilaginous joint connected by hyaline cartilage. A syndesmosis is a slightly movable articulation united by fibrous connective tissue forming an intraosseous membrane. Since both types of joints only have a virtual intraosseous space, infections of synchondrosis as well as syndesmosis manifest in most cases as osteomyelitis of the adjacent bones. Since the characteristics of these three types of arthritis are different, they will be presented separately.

Septic Arthritis of the Sternoclavicular Joint Introduction The sternoclavicular joint is a synovial joint connecting the axial skeleton with the upper extremity, and is involved in the movement of the upper extremities. It is composed of the anterior sternoclavicular ligament, interclavicular ligament, costoclavicular ligament, and two synovial cavities separated by a fibrocartilaginous disc [4–7]. Septic arthritis can occur by the hematogenous route, exogenously by direct injury, or by continuous spread from adjacent infections. Hematogenous infection of the sternoclavicular joint is rare compared to other joints, because this joint receives only a very small proportion of the Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

151

152

Bone and Joint Infections

total cardiac output. Its proximity to major vascular structures explains the exogenous infection in IV‐drug addicts using the internal jugular vein for injection [8]. Similarly, it may occur as a complication after subclavian venous catheterization [7]. In addition, due to the lack of substantial overlying soft tissue, exogenous infection by animal bites or scratches may occur. Continuous infection is transferred from adjacent lymph nodes or infections from the surrounding. Tuberculous sternoclavicular arthritis is reported to seed hematogenously during primary dissemination  [9,10]. However, since bilateral infection has been reported, continuous spread from sternoclavicular lymph nodes on both sides is also conceivable [10].

Epidemiology Sternoclavicular joint arthritis is a rare disease, especially in a population without risk factors. This may be one of the reasons why the diagnosis is often delayed [8]. Table 10.1 shows the prevalence of sternoclavicular arthritis in 1695 episodes of native joint arthritis published during the last five decades [11–18]. In many additional case series, the exact number of patients with sternoclavicular arthritis is not reported. The reported prevalence goes from 0.5 to 8.3% [12,18]. This 16‐fold difference is rather due to distinct risk factors in the populations than to the patients’ origin in different countries or the date of publication. Indeed, in the Australian study no patient with IV‐drug use has been reported, whereas in the Swiss study, 15% of the patients were IV‐drug addicts. This is an obvious explanation for the large difference in the two series (see risk factors). The lack of IV‐drug users despite HIV infections in the Thailand series may be due to obscured history taking. Overall, the mean prevalence of septic sternoclavicular arthritis is 3.4% in an unselected population [Table 10.1]. However, in IV‐drug addicts, axial joint involvement is observed in up to 68% of the cases [19]. In the study of Brancos et al. [20], 8/36 (22.2%) of the episodes of septic arthritis in IV‐drug users occurred in the sternoclavicular joint. TABLE 10.1.  Fraction of sternoclavicular arthritis in case series with septic arthritis.

Reference

Country

Sternoclavicular joint (SCJ) arthritis/ total arthritis

Argen et al., 1966 [11]

USA

5/60

8.3%

Morgan et al., 1996 [12]

Australia

1/191

0.5%

Kaandorp et al., 1997 [13]

NL

4/214

1.9%

Weston et al., 1999 [14]

UK

2/243

Ross et al., 2003 [15]

USA

2/108

Chanet et al., 2005 [16]

France

3/282

1.1%

Clerc et al., 2011 [17]

Switzerland

6/147

4.1%

Rodchuae et al., 2017 [18]

Thailand

33/450

7.3%

Total

4 continents

55/1695

3.2%

a  Only patients with pneumococcal arthritis. Sources: [11–18].

Percent patients with SCJ arthritis

0.8% a

1.9%

10  Septic Arthritis of Axial Joints

153

Microbiology S. aureus is the most frequent microorganism in all types of septic arthritis. In a cumulative statistic reported by Ross et al. [15], 1066/2407 (44%) of the cases with each type of septic arthritis were caused by S. aureus, 8% by pyogenic streptococci, and 6% by Streptococcus pneumoniae. The same first author reported the microorganisms which have been isolated in 176 cases of sternoclavicular septic arthritis [8] (Table 10.2). S. aureus is also by far the most frequent isolate (86/176=49%) in this type of arthritis. It is followed by Pseudomonas aeruginosa, which causes 10% of sternoclavicular joint infections. Interestingly, this microorganism causes only 1% of unselected types of arthritis [15]. This overrepresentation in patients with sternoclavicular arthritis is no more reported in more recent case series, presumably because sterile paraphernalia became widespread available for IV‐drug users [21,22]. Brucella melitensis is rare in most, but not all studies of sternoclavicular arthritis [8,21–23]. More frequently, it causes septic sacroiliitis [23].

Risk Factors According to Ross et al. [8], only about one‐quarter of the patients with septic sternoclavicular arthritis have no predisposing condition. However, even in the absence of risk factors, sternoclavicular septic arthritis must be considered in patients with pertinent signs and symptoms. Bar‐Natan et al. [24] reported a case series and review of 27 previously healthy adults. Since the publication of the comprehensive review by Ross et al. [8], the high prevalence of predisposing conditions has been confirmed in several case Table 10.2.  Microorganisms isolated in 176 cases of sternoclavicular septic arthritis (Data from Ross et al. [8]). Microorganism

Number of isolates

Stapylococcus aureus Pseudomonas aeruginosa Brucella melitensis Escherichia coli Group B streptococcus Mycobacterium tuberculosis Non‐specified streptococcus Streptococcus pneumoniae Anaerobes Group A streptococcus Haemophilus influeanzae type b Group G streptococcus Streptococcus milleri group Neisseria gnorrhoeae Other enteric Gram‐negative bacillia Miscellaneousb Polymicrobial

86 (49%) 18 (10%) 13 (7%) 8 (5%) 6 (3%) 6 (3%) 5 (3%) 4 (2%) 2 (1%) 2 (1%) 2 (1%) 2 (1%) 2 (1%) 2 (1%) 5 (3%) 7 (4%) 6 (3%)

a  One each of Acinetobacter anitratus, Burkholderia pseudomallei, Citrobacter diversus, Proteus mirabilis, and Serratia marcescens. b  One each of Candida albicans, Haemophilus aphrophilus, Mycobacterium avium complex, Pasteurella multocida, Cutibacterium acnes, Staphylococcus epidermidis, and Streptobacillus moniliformis.

154

Bone and Joint Infections

series [7,16,25–27]. The following risk factors have been reported in various frequency: intravenous drug use, diabetes mellitus, trauma, infected central line, chronic renal failure, hepatic dysfunction, alcohol abuse, corticosteroid, HIV infection, malignancy, and liver cirrhosis [8,25,26]. In addition, radiation therapy in the region of the sternoclavicular joint has also been reported as a risk factor. Chanet et al. [16] observed nine patients after radiation therapy for breast cancer. Six of them had septic shoulder arthritis and three suffered from sternoclavicular joint arthritis at the site of previous radiotherapy, after a median time interval of 16 years after irradiation. It is highly probable that there was a causal link between the two events, because the authors reported only 2/273 (0.65%) episodes of sternoclavicular arthritis without and 3/9 (33%) with previous radiotherapy. In the review of Ross et al. [8], 25 patients had a diagnosed primary focus, the most frequent being pneumonia (10/180), cellulitis (8/180), and endocarditis (3/180). Cinquetti et al.[27] reported sternoclavicular arthritis after Lemierre’s syndrome. It seems obvious that in this case, Fusobacterium necrophorum seeded continuously from the septic thrombosis of the jugular vein to the adjacent joint. Occasional cases caused by Pasteurella multocida were described in patients with licking dogs [28], and cats [29], respectively. Venous catheter‐related infections endanger the sternoclavicular joint by different pathogenic mechanisms. First, the infection may seed by the hematogenous route. Second, the joint may be directly inoculated during the puncture procedure. Finally, catheter‐ related infections may invade the adjacent capsule of the joint  [7]. In IV‐drug users, pathogenesis of infection is similar. These patients have an increased colonization by S. aureus, a higher risk of bacteremia, and occasionally attempt the direct puncture of the internal jugular vein between the heads of the sternocleidomastoid muscle. During this try, they may directly injure the adjacent joint with a contaminated needle.

Clinical and Laboratory Features In the largest series of Ross et al.[8], 73% of the patients were men. This male preponderance has also been observed in other studies going from 58% [21] up to 86% [25]. This sex imbalance is at least partly due to the overrepresentation of IV‐drug users in most series. Looking exclusively at this population, even 91% were of male sex  [8]. However, this cannot be the only explanation, since in Bar‐Natan et  al.’s  [24] series of 27 patients without underlying risk factors, 70% of the patients were male. According to the most comprehensive case series of Ross et al.[8], the leading symptom is chest pain (78%), which is sometimes localized in the shoulder (24%). The clinical spectrum of signs ranges from pure arthritis to periarticular inflammation, to osteomyelitis, to sinus tract, to frank abscess formation, and to mediastinitis [8,30]. Only 65% of the patients have >38°C fever, probably due to the fact that most patients use analgetic drugs. Functional impairment characterized by decreased range of shoulder motion is observed in less than 20% of the patients. The sternoclavicular joint is almost not distendable because of strong ligaments. Therefore, the leading clinical sign is a tender (90%), but not a swollen, joint (4%). There is a slight preponderance of right side arthritis (57%); in 5% both sternoclavicular joints are involved, and in 21% polyarticular septic arthritis is observed. This is especially typical for gonococcal arthritis. This etiology must be especially considered in patients with migratory tenosynovitis and oligoarthritis [31]. In patients with central venous catheter‐related infection, the diagnosis of sternoclavicular arthritis may be difficult. The typical leading sign is cellulitis around the insertion site [7].

10  Septic Arthritis of Axial Joints

155

Figure 10.1.  78‐year‐old man with monoclonal gammopathy of unknown significance presented with fever, cellulitis, and swelling on the left side of the upper thorax. Two blood cultures revealed S. aureus. A first CT scan revealed a large abscess from the neck to the left nipple. Despite surgical drainage and adequate IV antibiotics, the infection progressed. Left: The CT scan control performed five weeks after initial drainage and start of antibiotic therapy shows a pyarthros of the left sternocalvicular joint, osteomyelitis of the proximal left clavicula, and manubrium sterni with sequestrated bone, as well as a retrosternal abscess (arrow). Right: The reconstruction shows a dislocated joint (arrow) as well as erosions of the manubrium sterni and the first rib.

Since the sternoclavicular joint has only a minimal synovial space, arthritis rapidly progresses to osteomyelitis, by invasion of medial parts of the clavicle, sternum, and ribs. In Murga et al.’s study [32], 47% of the patients had intraoperatively confirmed osteomyelitis. Sometimes, pyogenic complications (chest wall abscess, mediastinitis) are the first signs of sternoclavicular arthritis. Overall, pyogenic complications such as abscess formation, septic thrombosis of the subclavian or internal jugular vein, mediastinitis, and empyema are frequent [33–35]. According to Ali et al. [1SCJ] [36] 8% of the patients suffer from a draining sinus tract and 78% from an abscess identified by imaging studies (Figure 10.1). In most series, mortality is 11Giga/l, and only 62% bacteremia. Thus, rapid diagnosis needs imaging procedures followed by joint puncture or biopsy.

Imaging Procedures Plain radiographs have a low sensitivity to reveal sternoclavicular arthritis. Three phase bone scan has a good sensitivity, but a low specificity for the diagnosis of osteomyelitis. In view of the frequency of pyogenic complications, CT scan or MRI should be performed in all cases [36]. 18F‐FDG PET‐CT is an alternative modality to diagnose septic

156

Bone and Joint Infections

sternoclavicular arthritis [37]. Gas within the joint space does not prove arthritis. It is also known as vacuum phenomenon in normal individuals [4]. MRI should be preferred to CT scan in case of suspicion of osteomyelitis.

Differential Diagnosis Septic sternoclavicular arthritis has a broad differential diagnosis including osteoarthritis, rheumatoid arthritis, ankylosing spondylitis, psoriasis, gout, and chondrocalcinosis [2]. If clinical symptoms of arthritis occur in the context of skin eruption, an aseptic neutrophilic dermatosis should be considered. This rare syndrome is labeled with the acronym SAPHO (synovitis, acne, palmoplantar pustulosis, hyperostosis, osteitis). In most cases (65–90%) with SAPHO, the sternoclavicular joint is involved [38].

Management The treatment aim is complete eradication of infection, relief of pain, and the recovery of function of the upper extremity. Since diagnosis is often delayed, this aim generally needs a combined noninvasive (antibiotics) and surgical approach. The spectrum of treatment goes from antibiotic therapy alone to combination with simple incision and drainage to invasive surgery, such as en‐bloc resection. Antibiotics should be withheld until appropriate microbiological diagnosis has been performed. The choice of antibiotics does not differ from the one in all other types of native arthritis (see Chapter  9). In general, treatment is started with intravenous antibiotics. If the infecting agent is susceptible to drugs with excellent bioavailability, oral treatment can be started after a few days [39]. The duration has not been tested in comparative studies. In general, a 6‐ week course is proposed. In case of pyogenic complications, which are not completely removed by surgery, a longer treatment may be required. Table 10.3 shows a classification that has been proposed by Abu Arab [25]. Patients with grade I–III can be treated either with antibiotics alone, or combined with incision Table 10.3.  Classification of sternoclavicular joint (SCJ) modified according to Abu Arab [25]. Grade I

Signs: Inflammation over SCJ with intact overlying skin, no systemic signs of infection Symptoms: Mild pain Imaging (X‐ray, CT, MRI): minimal swelling, no signs of osteomyelitis (intact clavicle, sternum, and first rib), minimal or no effusion at SCJ

Grade II

Signs: Moderate to large swelling of SCJ, + inflammation, + systemic signs of infection Symptoms: pain Imaging: moderate swelling, moderate to large effusion at SCJ, no signs of osteomyelitis (intact clavicle, sternum, and first rib)

Grade III

Signs and symptoms: Any of the criteria mentioned in Grade II SCJ Imaging: Any of the criteria mentioned in Grade II SCJ plus minimal radiological signs of osteomyelitis

Grade IV

Any of the clinical or imaging criteria of grade I–II plus any of the following: • Severe osteomyelitis • Sinus tract • Persistence or recurrence of infection

Grade V

Any of the clinical or imaging criteria of grade I–IV plus evidence of mediastinitis

10  Septic Arthritis of Axial Joints

157

and drainage in case of effusion or local abscesses [25,40]. Some patients with grade III may require debridement of necrotic bone. In patients with grade IV or V, sternoclavicular joint resection is generally needed [25,26,41–44]. A combined approach by a thoracic and an orthopedic surgeon may be favorable in case of resection arthroplasty. In case of large soft tissue defects, a plastic‐reconstructive surgeon should be consulted in addition. Ali et al. [36] showed that the risk of recurrence was lower in patients with myocutaneous flaps, compared to those treated with definitive wound vacuum therapy.

Key Points ●●

●● ●●

Low prevalence of 0.5–8.3% in general population with septic arthritis, but much more frequent (15–22%) in septic arthritis of IV‐drug addicts. Three‐quarters of patients are men. In most cases, a combined antibiotic and surgical management is needed.

Septic Arthritis of the Symphysis Pubis Introduction The symphysis pubis is a synchondrosis allowing minimal motion, but it undergoes big shearing forces in pregnant women, especially during delivery. In addition, athletes including runners, football, soccer, and ice hockey players suffer from repetitive traumatisms at the insertion of the adductor muscles, leading to tendinitis as well as inflammation and sclerosis of the symphysis pubis  [45,46]. The diagnosis of septic arthritis of the pubic symphysis is difficult, since the clinical pictures of non‐infectious osteitis pubis and septic arthritis of the symphysis are similar. In addition, in a joint lacking synovial fluid, arthritis cannot be clinically diagnosed. Therefore, it is important maintaining a high level of suspicion in case of typical risk factors, signs, and symptoms.

Epidemiology The incidence of septic arthritis of the pubic symphysis is very low. In 5 studies on septic arthritis reporting the localization of 1045 episodes, none reported a single case [12–15,17]. This may be at least in part an artifact, because some of the cases are probably reported as osteomyelitis and not as arthritis, since >95% of the cases have not only arthritis, but adjacent osteomyelitis. However, also in the osteomyelitis series of Waldvogel et al. [47], pelvic localization was rare, i.e. 3.8%. In IV‐drug addicts each type of septic axial arthritis is more frequent than in a general population of patients with septic arthritis. Indeed, in the review of Brancos et al. [20], reporting 217 episodes of septic arthritis in heroin addicts, 19 (8.8%) were localized in the symphysis pubis.

Microbiology According to Ross and Hu  [48], S. aureus causes one‐third of septic arthritis of the symphysis pubis. It is the predominant pathogen in athletes. One‐fourth is caused by Pseudomonas aeruginosa. This microorganism was cultured from even 87% of the patients

158

Bone and Joint Infections

with IV‐drug use. In the 1970s in the US, most IV‐drug addicts used pentazocine, an illicit drug, which was solved in non‐sterile water  [8]. The preponderance of this microorganism in drug addicts is explained by the fact that tap water is routinely contaminated with Pseudomonas aeruginosa [49]. Interestingly, Pseudomonas infection in IV‐drug users has been rarely observed in Europe, where most drug addicts inject heroin, which must be solved in acid. In the 1980s, lemon juice, occasionally containing Candida spp., was used for this purpose [50]. However, for at least 20 years sterile paraphernalia are offered in supervised drug‐injecting rooms. As a result, this pathogen disappeared in more recent case series of infections in IV‐drug users. Most patients with pelvic malignancy have anaerobic/aerobic mixed infection, mainly due to the presence of a sinus tract  [48]. After incontinence surgery, different enterobacteriaceae (Escherichia coli, Proteus spp., Klebsiella pneumonia, etc.), enterococci, or group B streptococci are observed. Other microorganisms such as Streptococcus pneumoniae, Salmonella spp., mycobacteria, or Brucella spp. are found in rare cases [48,51–53]. Bali et al. [54] reviewed nine cases with tuberculous pubic arthritis published during the last three decades. In addition, they cite studies published between 1888 and 1974, in which more than 100 cases suffering from tuberculosis of the symphysis pubis have been reported. Thus, before the availability of anti‐tuberculous drugs, mycobacteria apparently preferentially seeded in the symphysis.

Risk Factors Only a minority of patients has no predisposing condition. There are five main risk groups. Patients with previous incontinence surgery (Marshall‐Marchetti‐Krantz urethropexy) or pelvic malignancy (often with sinus tract) have the highest risk. Bouza et al. [53] reported three male patients suffering from septic arthritis of the symphysis pubis after implantation of an anti‐incontinence device. Interestingly, two of them had previous local irradiation for prostate cancer, which may also be a predisposing factor (confer sternoclavicular arthritis). In addition, athletes, IV‐drug users, and postpartal women are also at risk [48,55]. Interestingly, IV‐drug addicts are prone to all types of axial septic arthritis for unknown reasons.

Clinical and Laboratory Features The symptoms of so‐called osteitis pubis (sterile inflammation in athletes) and osteomyelitis of the pubic symphysis are similar [45]. According to Ross and Hu [48], as well as many case reports, pain is the leading symptom [52–64]. Due to the non‐specific symptoms, the mean delay until diagnosis was 29 days (1–180d) in a review of 100 cases [48]. With a high degree of suspicion, a straightforward workup allows rapid diagnosis in patients with a localized pubic pain. However, if the pain is localized in the groin (41%), thigh (15%), or hip (12%), more frequent alternative diagnoses have to be excluded. In individual case reports, pain has also been reported in the buttock [60] or in the testis [61]. Leading signs are pubic tenderness (88%), fever >38°C (74%), painful (waddling) gait (59%), and pain with hip motion (45%) [48]. Laboratory signs are either not sensitive or non‐specific, or both. Leukocytosis (>11 Giga/l) is observed in only one‐third of the patients. In contrast, increased sedimentation rate and/or C‐reactive protein have been mentioned in most case reports, but are also non‐specific.

10  Septic Arthritis of Axial Joints

159

Imaging Procedures Plain radiograph findings lag behind clinical symptoms by about two to four weeks. In the series of Ross and Hu [48], 24/76 (32%) of the pelvic radiographs were normal at time of the first examination. Later, signs of arthritis (symphyseal widening) or osteomyelitis (marginal erosions, bone destruction) can be observed. 99mTC‐MDP bone scan is positive within a few days, but has a low specificity [60]. It documents bone turnover, which is observed in each type of inflammation, including sterile osteitis. CT‐scan shows abnormalities at an early time‐point. It documents soft‐tissue inflammation (phlegmon, abscess or sinus tract) and symphyseal widening as signs of arthritis, and bone erosion and sequester as signs of osteomyelitis (Figure 10.2). The presence of bone sequesters differs osteomyelitis from athlete’s osteitis. Except during pregnancy, and in the immediate postpartum period, there is no gas observed within the normal pubic symphysis. Thus, this may be a sign of septic arthritis of the symphysis pubis. MRI is the most sensitive and specific imaging in patients with arthritis of the pubic symphysis. It shows inflammation of skin and muscles, intra‐ articular synovial fluid or abscesses, and early signs of osteomyelitis, which complicates arthritis in 97% of the patients [48]. PET‐CT is an alternative imaging method, if there is a contraindication for MRI [65].

Figure 10.2.  39‐year‐old woman with IV‐drug use (heroin and cocaine) since age 15. The patient suffered from HIV‐infection CDC B2 and chronic hepatitis C. Over five years, she had three episodes of Staphylococcus aureus endocarditis of the mitral valve. After a free interval of one year with occasional IV‐drug use, she suffered from severe pubic pain. Septic symphysitis pubis was suspected on a CT scan and proven with a biopsy of the symphysis. Bone biopsies and blood cultures showed growth of an identical S. epidermidis. At the end of a correct six‐week course of antibiotics, bone erosions (thin arrow), sequesters, and prepubic abscesses (thick arrow) were still visible.

160

Bone and Joint Infections

Differential Diagnosis The clinical differential diagnosis includes syndromes characterized by the same pain as pubic arthritis. It may imitate acute appendicitis (abdominal pain and fever), groin hernia (groin pain), slipped disk (backache and radiating pain to the thigh), prostatitis, and cystitis. Most of these entities can be easily excluded either by appropriate laboratory workup or by imaging procedures. Pathologic findings in different imaging procedures (plain radiograph, CT scan, bone scintigraphy) have a broad differential diagnosis including arthritis of the pubic symphysis, non‐infectious osteitis pubis, seronegative spondyloarthropathies, posttraumatic changes in the context of pregnancy or delivery, overuse stress injury in athletes (soccer, football, running, ice hockey, etc.), osteoporotic fracture, osteoarthritis and idiopathic hyperostosis, metabolic (hemochromatosis, etc.) and crystal‐induced diseases (gout, chondrocalcinosis), as well as malignancies. Osteitis pubis in athletes is generally hard to differentiate from septic pubic arthritis [65]. In addition, the presence of one diagnosis does not exclude the other, since osteitis pubis predisposes to subsequent infection. Consecutive appearance of both diagnoses has been [51] reported. Radiographic changes of the symphysis are frequent in soccer players. Harris and Murray [66] reported radiologic signs of osteitis pubis in 76% of players routinely examined. In addition, in 58% it correlated with symptoms such as groin and lower back pain. Patients with ankylosing spondylitis not only have signs of sacroiliitis, but frequently also have radiographic abnormalities of the pubic symphysis [67]. Less frequently, such changes are also observed in patients with psoriatic arthritis [68]. Calcification of the symphysis pubis is a rather common incidental manifestation of chondrocalcinosis. However, most of these episodes obviously occur asymptomatically, because there is only one single publication of clinical pseudogout in the pubic symphysis  [69]. Similarly, symptomatic tophaceous gout in the symphysis is also a very rare event [70].

Management All patients need antimicrobial therapy, as soon as the microbiologic testing shows growth of a relevant microorganism. Since septic arthritis can only be differentiated from osteitis pubis by the presence of a positive culture, antibiotics should be withheld until formal proof of infection. If blood cultures remain negative and cultures of the needle aspirates do not show any microbial growth, open debridement surgery for diagnostic purposes is needed before starting antimicrobial therapy. Since the diagnosis of septic pubic arthritis is often delayed, most patients suffer from concomitant osteomyelitis. In addition, in many patients retropubic abscesses are observed. Therefore, in at least half of patients, surgery is required. This includes incision and drainage of an abscess, debridement of dead bone, or both. The choice of the antimicrobial agent does not differ from the one in other type of native arthritis or osteomyelitis (see Chapters  9 and  18). In general, a duration of six weeks is suggested  [48]. Antimicrobial therapy is generally started intravenously. Treatment can be switched to the oral route, if drugs with excellent bioavailability are available [39].

10  Septic Arthritis of Axial Joints

161

Key Points ●●

●●

●●

Most patients have a predisposing condition, such as urogenital surgery, iv‐drug use, or participation in sports requiring forceful hip adduction (soccer, running, tennis). Pain and pubic tenderness are leading symptom and sign. With high degree of suspicion diagnostic delay can be avoided. Surgery is needed in case of chronic osteomyelitis or pyogenic complications.

Septic Arthritis of the Sacroiliac Joint Introduction The sacroiliac joint articulates between the sacrum and the innominate bone of the ­pelvis. Whereas the posterosuperior portion is a syndesmosis, the anterior portion is synovial  [71]. This joint is typically affected by inflammatory rheumatic diseases, such as ankylosing spondylitis or psoriatic spondylarthritis [72]. In addition, similar to the pubic symphysis, it undergoes mechanical stress during pregnancy, delivery, and heavy sport exposure  [73, 74]. Thus, the clinical differentiation between inflammatory rheumatic ­disease, mechanical stress, and septic arthritis may be difficult.

Epidemiology The incidence of septic sacroiliitis is low. In 6 studies reporting the localization of 963 episodes of septic arthritis, only 7 (0.73%) were localized in the sacroiliac joint [11–15, 17]. The incidence is about three times higher in children than in adults. In children, there is no sex preponderance, whereas in adults, 60–65% of the cases are females, probably due to the increased risk during pregnancy and postpartum  [75–79]. Adult septic sacroiliitis generally occurs in adolescents and during child‐bearing age [77]. Adults with septic sacroiliitis are about 10 years younger compared to adults with other types of septic arthritis [17, 75, 78, 79]. In IV‐drug users, the incidence of septic arthritis of the sacroiliac joint is very high. In a compiled statistic of 217 heroin addicts with arthritis, 67 (30.9%) were localized in the sacroiliac joint [20].

Microbiology According to a review of 326 cases with septic sacroiliitis, excluding mycobacteria and Brucella, the three most frequent causing agents were S. aureus (70%), streptococci (mainly Group A and B) (8.9%), and Pseudomonas aeruginosa (5.2%)  [80]. Any microorganism causing bacteremia can seed in the sacroiliac joint. As an example, in a young HIV‐infected patient, chlamydia DNA has been detected in a CT‐guided puncture from the sacroiliac joint, three weeks after an episode of urethritis [81]. Thus, detection of the infectious focus may give a hint regarding the microorganism causing septic sacroiliitis. In the review of Zimmermann et al. [80], 4.9% of the episodes were caused by Salmonella spp. Feldman [82] reviewed 24 Salmonella cases reported between 1977 and 2006. Most patients were adolescents (mean age 18.8 years), none of them had a known immunodeficiency, and none had sickle cell anemia. In endemic zones for Brucella

162

Bone and Joint Infections

(Mediterranean countries of Europe, Northern and Eastern Africa, Near East countries, India, Central Asia, Mexico, and Central and South America), brucellar sacroiliitis has to be considered, especially in bilateral indolent cases [83]. Tuberculous sacroiliitis has been observed in about 10% of the patients with bone and joint tuberculosis. It is also mainly reported from endemic areas. Gao et al. [84] reported 15 patients observed between 1997 and 2007 in China. In a very large original series from France, comprising 214 episodes of septic sacroiliitis, 65% were due to pyogenic microorganisms, 25% to Mycobacterium tuberculosis, and 10% to Brucella spp. [85].

Risk Factors IV‐drug use is the most frequently associated factor, reported in 14% of the cases. About 10% of children and 6.6% of adults suffering from septic sacroiliitis have a history of previous pelvic trauma [86]. In rare cases, septic sacroiliitis occurs during pregnancy or in the early postpartum period [78, 80]. In addition, sacroiliac joint steroid injection, which is an increasingly used intervention, can also result in septic sacroiliitis [87]. Since septic sacroiliitis is generally a hematogenous arthritis, some patients have a history of a primary infection. According to Vyskocil et  al.  [86], 7.2% of the patients suffered from skin infection, followed by upper (3.6%) and lower respiratory tract infection (1.8%), gynecologic (3%), and urinary tract (1.2%) infection.

Clinical and Laboratory Features The clinical diagnosis of septic sacroiliitis is difficult due to the nonspecific signs and symptoms. In addition, fever may be lacking, probably due to the frequent use of analgesics lowering the temperature. Therefore, confirmation of the diagnosis is often delayed [88]. In a literature review, fever and acute onset of pain was observed in 103/137 (75%) of the cases in which the type of presentation was given  [86]. Similarly, in Feldmann’s study  [85] reporting 138 patients with pyogenic sacroiliitis, 82% had a temperature of at least 38°C [85]. The classical triad of fever, low back pain, and difficulty in weight bearing is more frequently seen in pediatric (82%) than in adult (64%) patients [75]. The leading symptom is pain, described as low back, buttock, abdominal, pubic, hip, coxofemoral, thigh, and calf pain. The radiating pain can be explained by the fact that the first two sacral nerves (superior gluteal and obturator nerves) cross anterior to the sacroiliac joint [89]. Clinical signs can be summarized as fever, tenderness on pressure on the sacroiliac joint, tenderness on rectal palpation, pain exacerbation by weight‐bearing or moving the sacroiliac joint, painful limping, pain on hip mobilization, psoas sign, and painful abdomen [72, 75–78, 80–82, 84–101]. In addition, provocation maneuvers such as the Mennell test (also labeled as Gaenslen’s test in US literature) or FABER (Flexion, Abduction, External Rotation) test are generally positive. Leukocyte count is not a sensitive parameter. In a study by Wu et al. [75], only 46% had leukocytosis, whereas a C‐reactive protein (CRP) >60 mg/l was observed in 73% of the patients. The yield of positive blood cultures differs from 23 to 63% in four different case series [74, 75, 86, 88]. A low yield may be due to previous antibiotic therapy. At least, we observed eight non‐pretreated patients, all with positive blood cultures [101]. Synovial

10  Septic Arthritis of Axial Joints

163

fluid aspiration is rarely performed without evidence of fluid in an imaging procedure. However, the rate of positive synovial fluid aspiration is as high as 75%, if an abscess is visible on CT or MRI  [75]. If both blood cultures and synovial fluid do not show microbial growth, a biopsy of the sacroiliac joint should be performed before starting antimicrobial therapy.

Imaging Procedures Plain radiographs have a low sensitivity, and are usually normal until several weeks after infection [80, 86, 89, 102–104]. In contrast, 99mTc‐MDP bone scintigraphy is very sensitive for bone and joint infection and can be positive as early as two days after onset of symptoms [101]. After the first week, the majority is positive [89, 101]. However, bone scintigraphy is nonspecific, i.e. it shows the same pathology in all types of bony inflammation. In addition, due to the high bone turnover, there is a physiologic high uptake of technetium methyldiphosphonate in the sacroiliac joints. A pathologic accumulation of the tracer is easier to detect in unilateral arthritis. Thus, bone scintigraphy is still useful as screening test, if unilateral septic sacroiliitis is suspected. The hybrid procedure (single‐photon emission computed tomography/computed tomography: SPECT‐CT) allows a better localization of the inflammatory process [96]. However, this procedure is only more specific, if it is combined with labeled antigranulocyte antibodies or granulocyte scintigraphy [105]. Computed tomography allows the evaluation of soft‐tissue abnormalities, bone irregularities, joint erosions, and abscesses (Figure 10.3) [103]. Thus, it is especially useful as follow‐up examinations in patients with an unsatisfactory response to treatment, in order to look for pyogenic complications (e.g. M. iliacus abscess). As in all types of osteomyelitis, MRI is the most sensitive and specific imaging procedure. Bone marrow edema and inflammatory changes of the adjacent muscle tissue differentiate septic sacroiliitis from arthritis in patients with inflammatory rheumatoid disease [95, 106–108]. F‐18 FDG PET/CT hypermetabolic activity is also a sensitive method to detect inflammatory and septic sacroiliitis [109, 110]. However, its role in the diagnosis of septic sacroiliitis has to be confirmed.

Differential Diagnosis Since the sacroiliac joint can be affected by inflammatory rheumatic disease, trauma, and infection, the differential diagnosis is rather broad. It includes different types of seronegative spondyloarthropathies (e.g. psoriatic arthritis, ankylosing spondylarthropathy), crystal arthropathy, rheumatoid arthritis, familial Mediterranean fever, Behcet’s diseases, Whipple’s disease, hyperparathyroidism, and traumatic lesions [80, 83]. Occasionally, metastatic carcinoma or sarcoma may also mimic sacroiliitis [83]. In a French case series of 39 adults with septic sacroiliitis, the suspected clinical diagnosis at admission included lumbar disc herniation, vertebral osteomyelitis, mechanical low back pain, septic arthritis of the hip, inflammatory sacroiliitis, etc. [77]. According to a review of 191 cases published between 1929 and 1983, abdominal pain is a frequent symptom occurring in 24/191 (12.6%) of the cases  [111]. Thus, septic sacroiliitis is an imitator of the acute abdomen, which frequently led to laparotomy before CT scan and MRI were available.

164

Bone and Joint Infections

Figure 10.3.  49‐year‐old man, originally from Sri Lanka, with a three‐week history of pain in the left leg and sacrum. No trauma, no injections, intermittent fever, and shivering. At hospitalization 38.2°C, unable to stay on his left leg, pain in the region of the sacroiliac joint, and positive provocation maneuver (Mennell sign) on the left side. CRP 261 mg/l, leukocytes 11.9 G/l, growth of Staphylococcus aureus in 3/3 biopsies. The CT scan was performed after a three‐week history of symptoms at the time of hospitalization. It shows signs of spontaneous arthrodesis (thin arrow from dorsal) after septic arthritis of the left sacroiliac joint with osteomyelitis of the ileum (thick arrow) and a large abscess of the Musculus iliacus (thin arrow from ventral). Open surgical revision was performed.

Management Rapid intravenous antimicrobial therapy is the cornerstone of treatment. However, as in all types of bone and joint infection, antibiotics should only be started when infection is microbiologically documented. As an exception, in patients with a sepsis syndrome, empirical therapy should be started immediately after sampling of blood cultures and CT‐guided puncture of possible abscesses. Such an initial therapy has to be guided according to the most frequent spectrum of microorganisms, i.e. it is an educated guess. In IV‐drug addicts it should include the spectrum of S. aureus and Pseudomonas aeruginosa (e.g. piperacillin/tazobactam). In pregnant or postpartal women it should include Group B streptococci (e.g. ceftriaxone), and in patients with previous episodes of diarrhea, Salmonella spp. (e.g. ceftriaxone) should be covered by the empirical antibiotic. If the microorganism and its susceptibility are known, the choice of the antimicrobial agent does not differ from the one in other types of native arthritis or osteomyelitis (see Chapters 9 and 18). In general, a six‐week course is suggested [48]. Antimicrobial therapy is generally started intravenously. If possible, treatment can be switched to the oral route, if drugs with excellent bioavailability are available [39]. In case of rapid diagnosis, open surgical intervention is rarely required [101]. However, in the review of Ross et al. [48], the mean duration of symptoms before diagnosis was

10  Septic Arthritis of Axial Joints

165

29 days, and 55% required surgery. If imaging procedures reveal an abscess, CT‐guided drainage should be urgently performed for diagnostic and therapeutic reasons. Open surgery is needed in case of large abscesses, lack of response to antibiotics, sequesters, or late instability surgical debridement and arthrodesis may be required [80, 98, 100]. For this purpose, a minimally invasive technique has been described [98].

Key Points ●●

●●

●●

Mean age of adults with sacroilitis is about 10 years younger than in other types of septic arthritis. There is a female preponderance. Classical triad of fever, low back pain, and difficulty in weight bearing are present in only two‐thirds of the patients. Seronegative spondylarthropathy is the most frequent differential diagnosis.

References 1. Ross JJ. Septic Arthritis of Native Joints. Infect Dis Clin North Am. 2017;31(2):203–218. 2. Yood RA, Goldenberg DL. Sternoclavicular joint arthritis. Arthritis Rheum. 1980;23(2):232–239. 3. Renfree KJ, Wright TW. Anatomy and biomechanics of the acromioclavicular and sternoclavicular joints. Clin Sports Med. 2003;22(2):219–237. 4. Stark P, Jaramillo D. CT of the sternum. AJR Am J Roentgenol. 1986;147(1):72–77. 5. Restrepo CS, Martinez S, Lemos DF, et al. Imaging appearances of the sternum and sternoclavicular joints. Radiographics. 2009;29(3):839–859. 6. Barbaix E, Lapierre M, Van Roy P, et  al. The sternoclavicular joint: variants of the discus articularis. Clin Biomech (Bristol, Avon). 2000;15 Suppl 1:S3–7. 7. Pradhan C, Watson NF, Jagasia N, et al. Bilateral sternoclavicular joint septic arthritis secondary to indwelling central venous catheter: a case report. J Med Case Rep. 2008;2:131. 8. Ross JJ, Shamsuddin H. Sternoclavicular septic arthritis: review of 180 cases. Medicine (Baltimore). 2004;83(3):139–148. 9. Djiba B, Kane BS, Diallo MA, et  al. Tuberculosis arthritis of the sternoclavicular joint after uncomplicated falciparum malaria: a case report. Ann Clin Microbiol Antimicrob. 2017;16(1):44. 10. Pandita KK, Sharma R, Dogra S, et al. Bilateral sternoclavicular joint tubercular cold abscess. Ann Thorac Med. 2010;5(1):56–57. 11. Argen RJ, Wilson CH, Jr., Wood P. Suppurative arthritis. Clinical features of 42 cases. Arch Intern Med. 1966;117(5):661–666. 12. Morgan DS, Fisher D, Merianos A, et al. An 18 year clinical review of septic arthritis from tropical Australia. Epidemiol Infect. 1996;117(3):423–428. 13. Kaandorp CJ, Dinant HJ, van de Laar MA, et al. Incidence and sources of native and prosthetic joint infection: a community based prospective survey. Ann Rheum Dis. 1997;56(8): 470–475. 14. Weston VC, Jones AC, Bradbury N, et al. Clinical features and outcome of septic arthritis in a single UK Health District 1982‐1991. Ann Rheum Dis. 1999;58(4):214–219. 15. Ross JJ, Saltzman CL, Carling P, et al. Pneumococcal septic arthritis: review of 190 cases. Clin Infect Dis. 2003;36(3):319–327. 16. Chanet V, Soubrier M, Ristori JM, et al. Septic arthritis as a late complication of carcinoma of the breast. Rheumatology (Oxford). 2005;44(9):1157–1160.

166

Bone and Joint Infections

17. Clerc O, Prod’hom G, Greub G, et  al. Adult native septic arthritis: a review of 10 years of experience and lessons for empirical antibiotic therapy. J Antimicrob Chemother. 2011;66(5): 1168–1173. 18. Rodchuae M, Ruangpin C, Katchamart W. Clinical manifestations, treatment outcomes, and risk factors for sternoclavicular septic arthritis. Rheumatol Int. 2017;37(5):819–824. 19 Belzunegui J, Rodriguez‐Arrondo F, Gonzalez C, et al. Musculoskeletal infections in intravenous drug addicts: report of 34 cases with analysis of microbiological aspects and pathogenic mechanisms. Clin Exp Rheumatol. 2000;18(3):383–386. 20. Brancos MA, Peris P, Miro JM, et al. Septic arthritis in heroin addicts. Semin Arthritis Rheum. 1991;21(2):81–87. 21. Kang BS, Shim HS, Kwon WJ, et al. MRI findings for unilateral sternoclavicular arthritis: differentiation between infectious arthritis and spondyloarthritis. Skeletal Radiol. 2019;48(2): 259–266. 22. Von Glinski A, Yilmaz E, Rausch V, et al. Surgical management of sternoclavicular joint septic arthritis. J Clin Orthop Trauma. 2019;10(2):406–413. 23. Geyik MF, Gur A, Nas K, et al. Musculoskeletal involvement of brucellosis in different age groups: a study of 195 cases. Swiss Med Wkly. 2002;132(7‐8):98–105. 24. Bar‐Natan M, Salai M, Sidi Y, et al. Sternoclavicular infectious arthritis in previously healthy adults. Semin Arthritis Rheum. 2002;32(3):189–195. 25. Abu Arab W, Khadragui I, Echave V, et  al. Surgical management of sternoclavicular joint infection. Eur J Cardiothorac Surg. 2011;40(3):630–634. 26 Bakaeen FG, Huh J, Fagan SP, et al. Surgical treatment of sternoclavicular joint infections in cirrhotic patients. Am J Surg. 2008;195(1):130–133. 27. Cinquetti G, Banal F, Mohamed S, et al. [Rare complications of Lemierre’s syndrome: septic sternoclavicular joint arthritis and cavitating pneumonia]. Rev Med Interne. 2009;30(12): 1061–1063. 28. Blanchais A, Cormier G, Varin S, et al. [Pasteurella multocida septic oligoarthritis]. Med Mal Infect. 2010;40(2):123–125. 29. Nitsche JF, Vaughan JH, Williams G, et al. Septic sternoclavicular arthritis with Pasteurella multocida and Streptococcus sanguis. Arthritis Rheum. 1982;25(4):467–469. 30. Kachala SS, D’Souza DM, Teixeira‐Johnson L, et al. Surgical Management of Sternoclavicular Joint Infections. Ann Thorac Surg. 2016;101(6):2155–2160. 31. Guillot X, Delattre E, Prati C, et al. Destructive septic arthritis of the sternoclavicular joint due to Neisseria gonorrhoeae. Joint Bone Spine. 2012;79(5):519–520. 32. Murga A, Copeland H, Hargrove R, et al. Treatment for sternoclavicular joint infections: a multi‐institutional study. J Thorac Dis. 2017;9(6):1503–1508. 33. Pollack MS. Staphylococcal mediastinitis due to sternoclavicular pyarthrosis: CT appearance. J Comput Assist Tomogr. 1990;14(6):924–927. 34. Asnis DS, Dhaliwal GS. Bilateral sternoclavicular joint septic arthritis presenting as cutaneous abscesses. Clin Infect Dis. 1994;19(5):964–966. 35. Dhulkotia A, Asumu T, Solomon P. Breast abscess: a unique presentation as primary septic arthritis of the sternoclavicular joint. Breast J. 2005;11(6):525–526. 36. Ali B, Shetty A, Qeadan F, et al. Sternoclavicular joint infections: Improved outcomes with myocutaneous flaps. Semin Thorac Cardiovasc Surg. 2019. 37. Saad Aldin E, Sekar P, Saad Eddin Z, et  al. Incidental diagnosis of sternoclavicular septic arthritis with Moraxella nonliquefaciens. IDCases. 2018;12:44–46. 38. Earwaker JW, Cotten A. SAPHO: syndrome or concept? Imaging findings. Skeletal Radiol. 2003;32(6):311–327. 39. Li HK, Rombach I, Zambellas R, et al. Oral versus intravenous antibiotics for bone and joint infection. N Engl J Med. 2019;380(5):425–436. 40. Jang YR, Kim T, Kim MC, et al. Sternoclavicular septic arthritis caused by Staphylococcus aureus: excellent results from medical treatment and limited surgery. Infect Dis (Lond). 2019;51(9):694–700.

10  Septic Arthritis of Axial Joints

167

41. Joethy J, Lim CH, Koong HN, et al. Sternoclavicular joint infection: classification of resection defects and reconstructive algorithm. Arch Plast Surg. 2012;39(6):643–648. 42. Nusselt T, Klinger HM, Freche S, et al. Surgical management of sternoclavicular septic arthritis. Arch Orthop Trauma Surg. 2011;131(3):319–323. 43. Puri V, Meyers BF, Kreisel D, et al. Sternoclavicular joint infection: a comparison of two surgical approaches. Ann Thorac Surg. 2011;91(1):257–261. 44. Burkhart HM, Allen MS, Nichols FC, 3rd, et al. Results of en bloc resection for bronchogenic carcinoma with chest wall invasion. J Thorac Cardiovasc Surg. 2002;123(4):670–675. 45. Choi H, McCartney M, Best TM. Treatment of osteitis pubis and osteomyelitis of the pubic symphysis in athletes: a systematic review. Br J Sports Med. 2011;45(1):57–64. 46. Paajanen H, Ristolainen L, Turunen H, et al. Prevalence and etiological factors of sport‐related groin injuries in top‐level soccer compared to non‐contact sports. Arch Orthop Trauma Surg. 2011;131(2):261–266. 47. Waldvogel FA, Medoff G, Swartz MN. Osteomyelitis: a review of clinical features, therapeutic considerations and unusual aspects. N Engl J Med. 1970;282(4):198–206. 48. Ross JJ, Hu LT. Septic arthritis of the pubic symphysis: review of 100 cases. Medicine (Baltimore). 2003;82(5):340–345. 49. Fujitani S, Sun HY, Quintiliani R, et al. Pneumonia due to Pseudomonas aeruginosa: part II: antimicrobial resistance, pharmacodynamic concepts, and antibiotic therapy. Chest. 2011;139(5):1172–1185. 50. Miro JM, Puig de la Bellacasa J, Odds FC, et al. Systemic candidiasis in Spanish heroin addicts: a possible source of infection. J Infect Dis. 1987;156(5):857–858. 51. Pauli S, Willemsen P, Declerck K, et al. Osteomyelitis pubis versus osteitis pubis: a case presentation and review of the literature. Br J Sports Med. 2002;36(1):71–73. 52. Qassemyar Q, Smail A, Perignon D, et al. Cold abscess and Bazin’s indurated erythema associated to pubic symphysis tuberculosis. Med Mal Infect. 2010;40(9):552–554. 53. Bouza E, Winston DJ, Hewitt WL. Infectious osteitis pubis. Urology. 1978;12(6):663–669. 54. Bali K, Kumar V, Patel S, et al. Tuberculosis of symphysis pubis in a 17 year old male: a rare case presentation and review of literature. J Orthop Surg Res. 2010;5:63. 55. Cosma S, Borella F, Carosso A, et al. Osteomyelitis of the pubic symphysis caused by methicillin‐resistant Staphylococcus aureus after vaginal delivery: a case report and literature review. BMC Infect Dis. 2019;19(1):952. 56. Alaya Z, Osman W, Hassini L, et al. Osteopecilia associated with psoriatic arthritis. Pan Afr Med J. 2017;26:227. 57. Magarian GJ, Reuler JB. Septic arthritis and osteomyelitis of the symphysis pubis (osteitis pubis) from intravenous drug use. West J Med. 1985;142(5):691–694. 58. Brewer A. Acute pubic Osteomyelitis. Infect Disease Newsletter 12 (1). 1993. 59. Charles P, Ackermann F, Brousse C, et  al. Spontaneous streptococcal arthritis of the pubic symphysis. Rev Med Interne. 2011;32(7):e88–90. 60. Cheer K, Pearce S. Osteoarticular infection of the symphysis pubis and sacroiliac joints in active young sportsmen. BMJ. 2009;339:b5019. 61. Andole SN, Gupta S, Pelly M. Septic arthritis affecting pubic symphysis. BMJ Case Rep. 2011;2011. 62. To F, Tam P, Villanyi D. Septic arthritis of the pubic symphysis from pseudomonas aeruginosa: reconsidering traditional risk factors and symptoms in the elderly patient. BMJ Case Rep. 2012;2012. 63. Garcia‐Porrua C, Picallo JA, Gonzalez‐Gay MA. Osteitis pubis after Marshall‐Marchetti‐ Krantz urethropexy. Joint Bone Spine. 2003;70(1):61–63. 64. McPhee E, Eskander JP, Eskander MS, et al. Imaging in pelvic osteomyelitis: support for early magnetic resonance imaging. J Pediatr Orthop. 2007;27(8):903–909. 65. Ghislain L, Heylen A, Alexis F, et  al. Septic arthritis of the pubic symphysis: an atypical abdominal pain. Acta Clin Belg. 2015;70(1):46–49. 66. Harris NH, Murray RO. Lesions of the symphysis in athletes. Br Med J. 1974; 4(5938):211–214.

168

Bone and Joint Infections

67. Resnick D, Dwosh IL, Goergen TG, et al. Clinical and radiographic abnormalities in ankylosing spondylitis: a comparison of men and women. Radiology. 1976;119(2):293–297. 68. Maldonado‐Cocco JA, Porrini A, Garcia‐Mortoe O. Prevalence of sacroiliitis and ankylosing spondylitis in psoriasis patients. J Rheumatol. 1978;5(3):311–313. 69. Kenzaka T, Wakabayashi T, Morita Y. Acute crystal deposition arthritis of the pubic ­symphysis. BMJ Case Rep. 2013;2013. 70. Justiniano ME, Colmegna I, Gimenez CR, et  al. Tophaceous gout of the symphysis pubis. Arthritis Rheum. 2005;52(12):4052. 71. Sgambati E, Stecco A, Capaccioli L, et al. Morphometric analysis of the sacroiliac joint. Ital J Anat Embryol. 1997;102(1):33–38. 72. Ehrenfeld M. Spondyloarthropathies. Best Pract Res Clin Rheumatol. 2012;26(1):135–145. 73. Keriakos R, Bhatta SR, Morris F, et al. Pelvic girdle pain during pregnancy and puerperium. J Obstet Gynaecol. 2011;31(7):572–580. 74. Ruhe A, Bos T, Herbert A. Pain originating from the sacroiliac joint is a common non‐traumatic musculoskeletal complaint in elite inline‐speedskaters – an observational study. Chiropr Man Therap. 2012;20(1):5. 75. Wu MS, Chang SS, Lee SH, et al. Pyogenic sacroiliitis – a comparison between paediatric and adult patients. Rheumatology (Oxford). 2007;46(11):1684–1687. 76. Leroux J, Bernardini I, Grynberg L, et al. Pyogenic sacroiliitis in a 13‐month‐old child: A Case Report and Literature Review. Medicine (Baltimore). 2015;94(42):e1581. 77 Hermet M, Minichiello E, Flipo RM, et al. Infectious sacroiliitis: a retrospective, multicentre study of 39 adults. BMC Infect Dis. 2012;12:305. 78. Imagama T, Tokushige A, Sakka A, et  al. Postpartum pyogenic sacroiliitis with methicillin‐ resistant Staphylococcus aureus in a healthy adult: A case report and review of the literature. Taiwan J Obstet Gynecol. 2015;54(3):303–305. 79. Knipp D, Simeone FJ, Nelson SB, et al. Percutaneous CT‐guided sacroiliac joint sampling for infection: aspiration, biopsy, and technique. Skeletal Radiol. 2018;47(4):473–482. 80. Zimmermann B, 3rd, Mikolich DJ, Lally EV. Septic sacroiliitis. Semin Arthritis Rheum. 1996;26(3):592–604. 81 Rihl M, Wagner AD, Bakhsh KA, et al. Detection of chlamydial DNA in the inflamed sacroiliac joint of a patient with multiple infections. J Clin Rheumatol. 2009;15(4):195–197. 82. Feldman LS. Salmonella septic sacroiliitis: case report and review. Pediatr Infect Dis J. 2006;25(2):187–189. 83. Slobodin G, Rimar D, Boulman N, et  al. Acute sacroiliitis. Clin Rheumatol. 2016; 35(4):851–856. 84. Gao F, Kong XH, Tong XY, et al. Tuberculous sacroiliitis: a study of the diagnosis, therapy and medium‐term results of 15 cases. J Int Med Res. 2011;39(1):321–335. 85. Feldmann JL, Menkes CJ, Weill B, et al. Infectious sacroiliitis. Multicenter study of 214 cases. Rev Rhum Mal Osteoartic. 1981;48(1):83–91. 86. Vyskocil JJ, McIlroy MA, Brennan TA, et al. Pyogenic infection of the sacroiliac joint. Case reports and review of the literature. Medicine (Baltimore). 1991;70(3):188–197. 87. McHugh RC, Tiede JM, Weingarten TN. Clostridial sacroiliitis in a patient with fecal incontinence: a case report and review of the literature. Pain Physician. 2008;11(2):249–252. 88. Kucera T, Brtkova J, Sponer P, et al. Pyogenic sacroiliitis: diagnosis, management and clinical outcome. Skeletal Radiol. 2015;44(1):63–71. 89. Carlson SA, Jones JS. Pyogenic sacroiliitis. Am J Emerg Med. 1994;12(6):639–641. 90. Liu XQ, Li FC, Wang JW, et  al. Postpartum septic sacroiliitis misdiagnosed as sciatic ­neuropathy. Am J Med Sci. 2010;339(3):292–295. 91. Le Bars H, Lamini N, Brunet JF, et al. Sacroiliitis due to Kingella kingae in an adult: updates on this pathogen. Ann Biol Clin (Paris). 2010;68(3):341–345. 92. McKenna T, O’Brien K. Case report: group B streptococcal bacteremia and sacroiliitis after mid‐trimester dilation and evacuation. J Perinatol. 2009;29(9):643–645.

10  Septic Arthritis of Axial Joints

169

93. Bindal M, Krabak B. Acute bacterial sacroiliitis in an adult: a case report and review of the literature. Arch Phys Med Rehabil. 2007;88(10):1357–1359. 94. Cinar M, Sanal HT, Yilmaz S, et al. Radiological followup of the evolution of inflammatory process in sacroiliac joint with magnetic resonance imaging: a case with pyogenic sacroiliitis. Case Rep Rheumatol. 2012;2012:509136. 95. Bellussi A, Busi Rizzi E, Schinina V, et al. STIR sequence in infectious sacroiliitis in three patients. Clin Imaging. 2002;26(3):212–215. 96. Moros ML, Rodrigo C, Villacampa A, et al. Septic shock in pregnancy due to pyogenic sacroiliitis: a case report. J Med Case Rep. 2009;3:6505. 97. Roca B, Torres V. Pyomyositis of the iliacus muscle complicated with septic sacroiliitis. QJM. 2008;101(12):983–984. 98. Giannoudis PV, Tsiridis E. A minimally invasive technique for the treatment of pyogenic sacroiliitis. J Bone Joint Surg Br. 2007;89(1):112–114. 99. Kanakaris NK, Psarakis S, Chalidis B, et al. Management of pelvic instability secondary to chronic pyogenic sacroiliitis: case report. Surg Infect (Larchmt). 2009;10(4):353–358. 100. Ebraheim NA, Ramineni SK, Alla SR, et al. Sacroiliac joint fusion with fibular bone graft in patients with failed percutaneous iliosacral screw fixation. J Trauma. 2010;69(5):1226–1229. 101. Spoendlin M, Zimmerli W. Pyogenic sacroiliitis. Review of 8 personal cases and 200 cases from the literature. Schweiz Med Wochenschr. 1988;118(21):799–805. 102. Abid H, Chaabouni S, Frikha F, et al. Contribution of imaging in the diagnosis of infectious sacroiliitis: about 19 cases. Pan Afr Med J. 2014;17:171. 103. Ford LS, Ellis AM, Allen HW, et al. Osteomyelitis and pyogenic sacroiliitis: A difficult diagnosis. J Paediatr Child Health. 2004;40(5–6):317–319. 104. Brtalik D, Pariyadath M. A case report of infectious sacroiliitis in an adult presenting to the emergency department with inability to walk. J Emerg Med. 2017;52(3):e65–e68. 105 Rambaldi PF, Ambrosone L, Mansi L, et al. Immunoscintigraphy as an adjunct to diagnostic imaging in septic sacroiliitis. Report of a case. Clin Rheumatol. 1996;15(1):67–71. 106. Sturzenbecher A, Braun J, Paris S, et al. MR imaging of septic sacroiliitis. Skeletal Radiol. 2000;29(8):439–446. 107. Lambert RG, Bakker PA, van der Heijde D, et al. Defining active sacroiliitis on MRI for classification of axial spondyloarthritis: update by the ASAS MRI working group. Ann Rheum Dis. 2016;75(11):1958–1963. 108. Puhakka KB, Melsen F, Jurik AG, et al. MR imaging of the normal sacroiliac joint with correlation to histology. Skeletal Radiol. 2004;33(1):15–28. 109. Raynal M, Bouderraoui F, Ouichka R, et al. Performance of (18)F‐sodium fluoride positron emission tomography with computed tomography to assess inflammatory and structural sacroiliitis on magnetic resonance imaging and computed tomography, respectively, in axial spondyloarthritis. Arthritis Res Ther. 2019;21(1):119. 110. Ho CL, Wu WC, Chen S, et al. F‐18 FDG PET/CT in an adult case of group B streptococcal sacroiliitis. Clin Nucl Med. 2010;35(10):834–835. 111. Cohn SM, Schoetz DJ, Jr. Pyogenic sacroiliitis: another imitator of the acute abdomen. Surgery. 1986;100(1):95–98.

Chapter 11

Periprosthetic Joint Infection: General Aspects Werner Zimmerli

Introduction Prosthetic joints are used to replace articulations, which are damaged because of degen‑ eration, trauma, or inflammation. Since damaged cartilage can still not be replaced to a sufficient extent, arthroplasty remains the only treatment that is able to completely relieve pain and restore function. In case of an optimal treatment result, the prosthetic joint can be functionally used, similar to a native joint. Due to the increasing life expectancy, the number of patients with osteoarthritis needing an arthroplasty is steadily rising [1–3]. In contrast, the need for joint replacement in patients with rheumatoid arthritis is decreasing, since efficacious disease‐modifying drugs are widely available [4]. In parallel to the rise of primary arthroplasty, the need for revision arthroplasty is even higher [5]. Several complications can be observed after joint replacement, including mechanical problems such as luxation, heterotopic ossification [6], aseptic loosening because of wear particles [7], or insufficient adaptation of the implant shape to the bone [8]. The most feared complication, however, is periprosthetic joint infection (PJI), because it may lead to the loss of the device [9–11]. Despite considerable effort for prevention, the incidence of PJI did not significantly change between 2005 and 2015  [12]. Thus, there are considerable economic consequences. Kurtz et al. [13] projected a total cost of US$1.6 billion for surgical procedures in patients with PJI after total hip and total knee arthroplasty in 2020. Thus, it is of paramount interest for patients and society to avoid PJI, or, if it occurs, to manage this complication in an optimal way. Regardless of the type of device, implant‐associated infections have some common features: They are highly susceptible to infection, and cannot be cured with antibiotics alone [9–11, 14–16]. In addition, prosthetic devices are endangered for infection not only during implantation, but as long as they remain in the body  [17–19]. In this chapter, common features of PJI will be presented. Specific problems that are associated with

Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition. Edited by Werner Zimmerli. © 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.

171

172

Bone and Joint Infections

each type of artificial joint are discussed in separate chapters dealing with the different joint replacements (see Chapters 12–15).

Definition There is no unanimously accepted gold standard for the diagnosis of PJI. Therefore, various diagnostic criteria have been used in different publications [9, 11, 20]. The two most commonly used are the Infectious Disease Society of America (IDSA) criteria and the Musculoskeletal Infection Society (MSIS) criteria [11, 20]. Both definitions are not ideal. The IDSA criteria miss important criteria, namely synovial cell counts and neutrophil ratio, whereas the MSIS criteria are not sensitive enough. Therefore, the IDSA criteria were modified accordingly  [21, 22]. In addition, the MSIS criteria have been modified at the recent consensus conference. The new criteria were shown to have a good sensitivity (97.7%) and specificity (99.5%), as shown in a cohort of 222 patients with PJI and 200 patients with aseptic revision [23]. For clinical studies, the definition should have a very high specificity for a meaningful comparison of published results. In contrast, in clinical practice, the sensitivity should be as high as possible, in order not to miss any PJI. If the diagnosis is delayed by 3–4 weeks, cure with implant retention has a very low chance [9–11]. Therefore, in such cases, implant removal is generally needed. As a consequence, even if a patient does not strictly fulfill the definition of PJI, diagnostic surgical debridement should be considered. The rational diagnostic approach in different types of bone and joint infections is presented in a sepa‑ rate chapter (see Chapter 2). Surgical site infections can be classified as superficial incisional, deep incisional, and organ/space infections [24]. However, this classification is not reliable after joint replace‑ ment. Superficial infection rapidly progress to deep infection, and the differentiation is clinically not possible [25]. With the wrong diagnosis of superficial surgical site infec‑ tion, a short course of antibiotics is given, which typically delays the diagnosis of PJI.

Classification Traditionally, PJIs are classified as early (2 years after surgery) [10, 11]. Early and delayed infec‑ tions are mainly exogenously acquired in the perioperative period, whereas most late PJIs are hematogenously acquired. For clinical purposes, a classification considering the surgi‑ cal treatment concepts is more useful. Table  11.1 shows the three types of PJIs. Acute hematogenous PJIs of less than 3 weeks’ duration and early postinterventional PJIs (3 weeks and are beyond the early postinterventional period

more staphylococci were required to get a subcutaneous abscess in the absence of a foreign body. We reproduced the same phenomenon in a guinea pig model with subcutaneous tis‑ sue cages [14, 15, 30]. These experimental observations point toward a locally acquired granulocyte defect, since staphylococcal infection is controlled by phagocytes. Once established, device‐related infections are difficult to eradicate. This indicates a locally impaired host defense on the one side and a phenotypic resistance of adhering microorganisms on the other side. The reason for this is an obvious unfavorable interac‑ tion between the local defense mechanism (e.g., granulocytes, complement) with the implanted foreign body, and a transition from the susceptible planktonic microorganism to a resistant biofilm. Interestingly, the type of material plays only a minor role regarding the susceptibility of a foreign body to infection  [31]. Ha et  al.  [32] demonstrated that biofilm‐forming Staphylococcus epidermidis adhere to a higher degree on pure titanium than on stainless steel. Nevertheless, this difference could not be reproduced in vivo. The reason for this paradox could be an immediate coating of the implant by host proteins, which are more relevant for bacterial adherence than the type of material. The main argument for this mechanism is the observation that fibronectin and other proteins act as receptors for staphylococci [33, 34].

The Role of the Host Innate or nonspecific host defense mechanisms are responsible for rapid and efficacious elimination of microorganisms, as soon as the device is placed in the body [16]. As a first step, microorganisms have to be opsonized for rapid ingestion by granulocytes or mono‑ nuclear leukocytes. This process involves nonspecific (complement, bacterial remnants) and specific (antibodies) soluble components in the humoral phase and intact corre‑ sponding receptors on phagocytes. If one component of this process is impaired, the susceptibility to infection is increased. Various possible mechanisms for the impaired bacterial clearance have been hypothesized. In addition, the paradox of microbial ­persistence in the presence of abundant granulocytes around the implant has been studied. Interaction of the Implant with Granulocytes The role of granulocytes is killing microbes and elimination of foreign bodies by phago‑ cytosis (Figure 11.1) [16]. If the foreign material is too large for phagocytosis, granulo‑ cytes interact with the nonphagocytosable surface, a process that is called frustrated phagocytosis  [35, 36]. This concept has been tested in an animal model  [14, 15]. The guinea pig tissue cage model perfectly simulates the clinical situation, that is, very low

174

Bone and Joint Infections

Figure 11.1.  Scanning electron micrograph (SEM) of an experimental implant‐associated infection. Sampling for SEM was performed 3 hours after infection. The picture shows two granulocytes facing an aggregate of Staphylococcus aureus. The irregular surface of S. aureus shows exopolysaccharides of the young biofilm. Reprinted with permission from Zimmerli and Sendi [16].

numbers of microorganisms lead to permanent implant‐associated infection, which never spontaneously heals [9, 10]. Even so‐called apathogenic bacteria such as Cutibacterium acnes or S. epidermidis cause infection in this model [37, 38]. Granulocytes purified from the interstitial fluid accumulating around subcutaneous implants indeed revealed a severe defect in ingestion, staphylococcal killing, and superoxide production [14, 15]. In vitro experiments indicate that this defect is due to the interaction with the non‐phagocytosa‑ ble surface. Granulocytes partially degranulate during interaction with the implant [15]. Since the liberated granulocytes also contain collagenase, this phenomenon may also be responsible for loosening of the device during infection [39]. Interaction of Wear Particles with Phagocytes After arthroplasty, wear particles are produced in variable amounts depending on the biomechanical situation. Abundant wear particles are supposed to be a risk factor for infection [40]. Therefore, not only the implant, but also wear particles may interact with granulocytes. Indeed, Bernard et al. [41, 42] described an impaired bactericidal activity of neutrophils after interaction with wear particles in vitro. Similarly, in the tissue surround‑ ing an implant with wear particles, cytokines are liberated by local macrophages  [43]. Some of these cytokines, such as M‐CSF and TGF‐alpha, directly stimulate osteoclas‑ togenesis, which favors implant loosening by bone resorption.

The Role of the Microorganism (Biofilm) Microorganisms that are not immediately eliminated during insertion of a device rapidly adhere to the implant and resist elimination by host defense mechanisms [44]. Therefore,

11  Periprosthetic Joint Infection: General Aspects

175

preventive measures are of paramount importance in implant surgery. Adherence to the surface involves nonspecific physical factors (e.g., surface tension, hydrophobicity, and electrostatic interaction) or specific adhesins such as fibronectin. This initial process is followed by biofilm formation, which is mediated in part by the polysaccharide intercel‑ lular adhesion encoded by the intercellular adhesion (ica) operon [45]. Within biofilms, microbes are enclosed in a polymeric matrix and develop into organized, complex ­communities, resembling multicellular organisms. At high microbial density, so‐called quorum‐sensing genes are activated in order to control the size of the biofilm [46]. Biofilm bacteria are up to 1000‐fold more resistant to antibiotics than planktonic bacteria [47]. This explains the persistence of implant‐associated infection once it is established. There is a transformation of the biofilm over time. Clinically, it can be observed that in case of implant retention, the chance of cure dramatically falls from about 80–90% to 30–60%, if treatment is started later than 3–4 weeks after infection [48‑54].

The Route of Infection Inoculation of PJI occurs either by the exogenous or by the hematogenous route [9, 10, 55]. Exogenous infections are mainly acquired during the perioperative period. They manifest themselves during the first 2 years after surgery. The less virulent the infecting agent, the more delayed the infection. As an example, the diagnosis of PJI due to Cutibacterium acnes is often delayed by months or even years [56, 57]. Microorganisms reach the implant not only during but also after surgery, as long as the wound is not completely dry. Microorganisms penetrate along drainage tubes or directly through the wound. The risk is especially high in patients with large hematoma. Thus, the primary aim in the early postoperative period is rapid wound healing. Therefore, in case of wound healing disturbance due to high volume secretion, diagnostic, and therapeutic debridement surgery should not be delayed. Rarely, exogenous infection occurs after arthrocentesis or after spontaneous or traumatic skin perforation from outside or from inside the device. Exogenous PJI is more frequent in joints with poor soft tissue coverage such as knee, elbow, or ankle arthroplasty. Indeed, 85% of periprosthetic ankle joint [58], but only 57% of hip joint infections occurred by the exogenous route  [51]. Looking exclusively at patients with S. aureus PJI, the difference was even more impressive. PJI occurred by the exogenous route in 45% of the knee, but only 22% of the hip arthroplasties [55]. Hematogenous PJI are acquired via bloodstream at any time after surgery. The incidence rate of knee and hip PJI is estimated as 5.9 episodes per 1000 joint‐years during the first 2 years, and 2.3 episodes per 1000 joint‐years thereafter [59]. This difference in susceptibility reflects an early predominance of perioperative infections on the one hand, and higher initial risk for hematogenous PJI on the other side. Due to the earlier‐ mentioned locally acquired granulocyte defect around the implant, the device is a locus minoris resistentiae, which is prone to hematogenous seeding. Indeed, Blomgren et al. [60, 61] showed that knee prostheses could be infected by the hematogenous route in a rabbit model. We quantified the risk for bacterial seeding in the guinea pig tissue cage infection model [62]. With a bacteremia of 1000 CFU S. aureus per ml blood, 42% subcutaneous implants could be selectively infected. With lower bacterial density in the bloodstream, no extravascular devices were infected. With a higher bacterial load, bacterial seeding was not selective, because not only implants, but also different organs were infected. Taken together, implants are favorite sites of bacterial seeding at the high bacterial density occurring during S. aureus bacteremia (>1000 CFU/ml blood).

176

Bone and Joint Infections

Laboratory Investigation Several laboratory parameters have been used to screen and to confirm PJI. Some of them are not sensitive enough and can therefore only be used as supportive arguments for PJI. In this chapter, an overview of different diagnostic markers is given. In Chapter 2, the diagnostic approach to PJI, and in Chapter 4 novel microbiological tests are presented in more detail.

Infectious Parameters in Blood Leukocyte counts are not useful for the diagnosis of PJI. In a meta‐analysis, the pooled sensitivity was only 45%, and the specificity 87% [63]. ESR and CRP have a much better sensitivity. In a systematic review including 30 studies with various cutoffs for ESR and CRP, the sensitivities are 75% and 88%, respectively, and the specificities 70% and 91%, respectively  [63]. In a recent single‐center study, CRP‐values were particularly low in patients with chronic PJI and those with low‐grade microorganisms such as Cutibacterium spp. and coagulase‐negative staphylococci [64]. In another study, in 32% of 73 consecutive patients with PJI, CRP levels were 0.85 μg/l) was not better than the one of CRP (>10 mg/l) (71% vs 68%); and as could be expected, the specificity of D‐dimer was even lower (80% vs 93%).

Synovial Cells The threshold of leukocyte counts in synovial fluid for the diagnosis of PJI is much lower than for septic native joint arthritis. The interpretation is different according to the underlying disease (degenerative or inflammatory arthritis), the localization of the joint, and the time interval after implantation. Since there are only studies performed in patients with hip or knee arthroplasty, results from arthrocenteses in other joints cannot be compared with published data. In three studies, patients with rheumatoid arthritis, those with joint hemorrhage, and those in the early postoperative period have been excluded [71– 73]. In patients after hip arthroplasty, Schinsky et al. [72] reported an optimal cutoff at 4200 leukocytes per μl and/or 80% neutrophil fraction. In two studies testing patients with knee arthroplasty, the optimal cutoff values were at 1700 and 1100 leukocytes per μl,

11  Periprosthetic Joint Infection: General Aspects

177

respectively, and the corresponding neutrophil fraction at 65 and 64%, respectively [71, 73]. According to Cipriano et al. [74], the optimal cutoff values were similar in patients with and without underlying inflammatory arthritis. In contrast, in patients with arthrocentesis within 6 weeks after total knee arthroplasty, the optimal cutoff values were much higher, namely, 27,800 leukocytes per μl, and a neutrophil fraction of 89% [75].

Synovial Fluid Leukocyte Esterase Leukocyte esterase is an enzyme secreted by activated granulocytes. Therefore, it is a marker for stimulated and lysed granulocytes in synovia. In a recent metanalysis, it had sensitivity of 0.90 (95% CI 0.76–0.96) and a specificity of 0.97 (95% CI 0.95–0.98) for the diagnosis of PJI [76]. Thus, this test can be used in centers where the cell counts are not available.

Histopathology Intraoperative frozen sections should only be used in centers with experienced pathologists, who are able to differentiate between mechanical failure and infection. Biopsies, which are sampled during surgery, should be divided in two parts: one for microbiology and the other for conventional histopathology. Comparing pairs of biopsies allows a better interpretation of culture results (contamination). In addition, in case of negative culture, the presence of granulocytes in biopsy specimens indicates culture‐negative PJI. According to a histopathological consensus classification, the best discriminative threshold for the diagnosis of PJI is the presence of 23 neutrophil granulocytes in a total of ten high‐power fields [77].

Culture Swab cultures have a low sensitivity and should therefore be avoided. Culture of synovial fluid has a sensitivity of about 85% and a specificity of at least 95% [78]. The sensitivity is better when using polymerase chain reaction (PCR) [79] or by culturing synovial fluid in blood culture flasks [80]. For the diagnosis of PJI during surgery, at least three, but optimally six biopsies, should be obtained [11].

Sonication Culture and Molecular Diagnosis Microorganisms persist as a biofilm on the implant. Sonication has been used, in order to detach microbes from the surface. In many centers, this technique is used for examination of explanted prostheses or modular parts. Since sonication may harm the viability of microorganisms, only evaluated technical protocols should be used [81, 82]. In the study of Trampuz et al. [81], sonication was significantly more sensitive only in patients treated with antibiotics within 2 weeks before sampling (75% versus 45%, P 35 or 30%) during documented Staphylococcus aureus bacteremia, as shown in an animal model and in observational studies [12–14]. In contrast, it is relatively low (< 2%) during exposure to a remote infection. In the Geneva cohort, only 1 PJI occurred per 79 exposures to remote infections (1.25%) [15].

Microbiology Almost all microorganisms, including Mycobacterium tuberculosis [16], nontuberculous mycobacteria [17], Mycoplasma [18], Legionella [19], and fungal agents [20], have been reported in PJI (see Chapter 3). The relative frequency of the microorganisms differs in different joints. However, in all types of PJI, staphylococci predominate. Table  12.1 shows the microorganisms in PJI hip and knee infection according to their pathogenesis (all types versus hematogenous PJI). Data from seven studies with a non‐selected population, reporting only cases from a defined joint, and giving the type of microorganisms in sufficient detail, are summarized [21–27]. After THA, the predominant microorganism responsible for PJI was S. aureus, causing 35% of the episodes. In TKA PJI, S. aureus and coagulase‐negative staphylococci (CNS) were about equally frequent, causing 31.8% and 27.1% of the episodes, respectively. For other microorganisms, there were no remarkable differences between the two different joints. Polymicrobial infections were observed in 2.3% of the patients with THA and 9.1% of those with TKA. In patients selected for hematogenous PJI, a predominance not only of S. aureus, but also of Streptococcus spp. and Gram‐negative bacilli, was observed [26,28]. Interestingly, Gram‐negative bacilli more frequently caused PJI after THA than after TKA for unknown reasons. Low‐virulence microorganisms, such as coagulase‐negative staphylococci or Cutibacterium spp., rarely cause hematogenous PJI, as could be expected.

12  Periprosthetic Joint Infection after Total Hip and Knee Arthroplasty

189

Table 12.1.  Microorganisms in periprosthetic hip and knee infection according to pathogenesis.

Microorganism

Total hip arthroplastya All types n=389

Total hip arthroplastyb Hematogenous n=45

Total knee arthroplastyc All types n=645

Total knee arthroplastyd Hematogenous n=158

Staphylococcus aureus

35%

37%

31.8%

58.9%

Coagulase‐negative staphylococci

23.6%

6.5%

27.1%

5.1%

Streptococcus spp.

8.7%

26.0%

7.8%

24.1%

Enterococccus spp.

7.5%

10.9%

6.2%

4.4%

Cutibacterium spp.

0.3%

0%

1.9%

0%

Gramnegative bacilli

9.8%

15.2%

5.9%

6.3%

Miscellaneous

10%

2.2%

1.9%

0.6%

Polymicrobial

2.3%

No growth

2.8%

9.1% 2.2%

8.3%

0.6%

 Giulieri et al. [21], Schinsky et al. [22], Gundoft et al. [23].  Rakow et al. [28]. c  Trampuz et al. [24], Laffer et al.[25], Stefansdottir et al. [26], Holmberg et al. [27]. d  Rakow et al. [28], Stefansdottir et al. [26]. a

b

Clinical Features The clinical presentation of PJI depends on the pathophysiology (exogenous versus hematogenous), the type of joint, and the duration of infection. The hallmarks of acute exogenous PJI are local signs of inflammation, such as skin erythema, hyperthermia, wound healing disturbance, drainage of fluid through the open wound, and purulent discharge [29,30]. The majority of patients with exogenous PJI have no fever, i.e. a temperature