- BONE AND JOINT INFECTIONS _ from microbiology to diagnostics and treatment -WILEY-BLACKWELL (2021)

Prévia do material em texto

<p>Bone and Joint Infections</p><p>This book is dedicated to my wife Annelies</p><p>Bone and Joint Infections</p><p>From Microbiology to Diagnostics</p><p>and Treatment</p><p>Editor</p><p>Werner Zimmerli</p><p>Second Edition</p><p>This edition first published 2021</p><p>© 2021 John Wiley & Sons Ltd</p><p>Edition History</p><p>First edition © 2015 by John Wiley & Sons, Inc.</p><p>All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form</p><p>or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to</p><p>obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.</p><p>The right of Werner Zimmerli to be identified as the author of the editorial material in this work has been asserted in accord-</p><p>ance with law.</p><p>Registered Offices</p><p>John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA</p><p>John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK</p><p>Editorial Office</p><p>9600 Garsington Road, Oxford, OX4 2DQ, UK</p><p>The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK</p><p>Boschstr. 12, 69469 Weinheim, Germany</p><p>1 Fusionopolis Walk, #07‐01 Solaris South Tower, Singapore 138628</p><p>For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.</p><p>wiley.com.</p><p>Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in</p><p>standard print versions of this book may not be available in other formats.</p><p>Limit of Liability/Disclaimer of Warranty</p><p>The contents of this work are intended to further general scientific research, understanding, and discussion only and are not</p><p>intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians</p><p>for any particular patient. In view of ongoing research, equipment modifications, changes in governmental regulations, and the</p><p>constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate</p><p>the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things,</p><p>any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors</p><p>have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or</p><p>completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied</p><p>warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales</p><p>representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or</p><p>product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher</p><p>and authors endorse the information or services the organization, website, or product may provide or recommendations it may</p><p>make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice</p><p>and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate.</p><p>Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was</p><p>written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial</p><p>damages, including but not limited to special, incidental, consequential, or other damages.</p><p>Library of Congress Cataloging‐in‐Publication Data</p><p>Name: Zimmerli, Werner, 1948– editor.</p><p>Title: Bone and joint infections : from microbiology to diagnostics and</p><p>treatment / editor, Werner Zimmerli.</p><p>Other titles: Bone and joint infections (Zimmerli)</p><p>Description: Second edition. | Hoboken, NJ, USA : Wiley-Blackwell, 2021. |</p><p>Includes bibliographical references and index. | Description based on</p><p>print version record and CIP data provided by publisher; resource not</p><p>viewed.</p><p>Identifiers: LCCN 2020047588 (print) | LCCN 2020047589 (ebook) | ISBN</p><p>9781119720683 (epub) | ISBN 9781119720669 (adobe pdf) | ISBN 9781119720652</p><p>(hardback) | ISBN 9781119720652q(hardback) | ISBN 9781119720669q(adobe</p><p>pdf) | ISBN 9781119720683q(epub)</p><p>Subjects: | MESH: Osteomyelitis | Arthritis, Infectious</p><p>Classification: LCC RC931.O7 (ebook) | LCC RC931.O7 (print) | NLM WE 251 |</p><p>DDC 616.7/15–dc23</p><p>LC record available at https://lccn.loc.gov/2020047588</p><p>LC record available at https://lccn.loc.gov/2020047589</p><p>Cover Design: Wiley</p><p>Cover Image: Courtesy of Werner Zimmerli</p><p>Set in 10/11.5pts Times New Roman MT Std by SPi Global, Pondicherry, India</p><p>10 9 8 7 6 5 4 3 2 1</p><p>http://www.wiley.com/go/permissions</p><p>http://www.wiley.com/</p><p>http://www.wiley.com/</p><p>v</p><p>List of Contributors xii</p><p>Preface to the Second Edition xvi</p><p>Foreword to the First Edition xvii</p><p>Acknowledgments xix</p><p>Chapter 1 Introduction 1</p><p>Werner Zimmerli</p><p>Chapter 2 Diagnostic Approach in Bone and Joint Infections 5</p><p>Nora Renz, Donara Margaryan, and Andrej Trampuz</p><p>Introduction 5</p><p>Common Microorganisms Causing Bone and Joint Infection 7</p><p>Diagnostic Approach in Spinal Infection 8</p><p>Diagnostic Approach in Bone Fixation Device‐Associated Infection 10</p><p>Diagnostic Approach in Native Arthritis and Infections after</p><p>Anterior Cruciate Ligament Repair (ACL‐R) 12</p><p>Diagnostic Approach in Periprosthetic Joint Infections (PJI) 15</p><p>Key Points 17</p><p>References 17</p><p>Chapter 3 Unusual Microorganisms in Periprosthetic Joint Infection 21</p><p>Camelia Marculescu and Werner Zimmerli</p><p>Introduction 21</p><p>Gram‐Positive Microorganisms 21</p><p>Gram‐Negative Microorganisms 26</p><p>Zoonotic Microorganisms 29</p><p>Anaerobic Microorganisms 34</p><p>Mycobacteria 35</p><p>Other Microorganisms 38</p><p>Key Points 40</p><p>References 40</p><p>Contents</p><p>vi Contents</p><p>Chapter 4 Identification of Pathogens in Bone and Joint Infections</p><p>by Non‐Culture Techniques 51</p><p>Maria Eugenia Portillo and Stéphane Corvec</p><p>Introduction 51</p><p>Broad‐Range PCR 51</p><p>Targeted PCR 53</p><p>Multiplex PCR 54</p><p>Next‐Generation Sequencing Approach 55</p><p>Mass Spectrometry‐Based Methods 58</p><p>Key Points 61</p><p>References 61</p><p>Chapter 5 Bacteriophages for Treatment of Biofilm Infections 65</p><p>Mercedes Gonzalez‐Moreno, Paula Morovic, Tamta Tkhilaishvili,</p><p>and Andrej Trampuz</p><p>History of Bacteriophage Use in Human Infections 65</p><p>Principles of Bacteriophage Therapy 66</p><p>Activity of Bacteriophages against Bacterial Biofilms and Persisters 69</p><p>Bacteriophage Susceptibility Testing: the Phagogram 71</p><p>Experimental and Clinical Evidence with Bacteriophage Treatment 72</p><p>Local Delivery and Systemic Bacteriophage Application 74</p><p>Outlook and Future Perspectives 76</p><p>Key Points 77</p><p>References 78</p><p>Chapter 6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 81</p><p>Cornelia B. Landersdorfer, Jürgen B. Bulitta, Roger L. Nation,</p><p>and Fritz Sörgel</p><p>Pharmacokinetics 81</p><p>Bone Sample Preparation and Analysis 82</p><p>Pharmacokinetic Sampling and Data Analysis 83</p><p>Penetration of Antibiotics into Bone 84</p><p>Pharmacodynamics and Monte Carlo Simulations 93</p><p>Conclusions 95</p><p>Key Points 96</p><p>References 96</p><p>Chapter 7 Preclinical Models of Infection in Bone and Joint Surgery 99</p><p>Caroline Constant, Lorenzo Calabro, Willem‐Jan Metsemakers,</p><p>R. Geoff Richards, and T. Fintan Moriarty</p><p>Introduction 99</p><p>Influence of Species in Preclinical Models of Bone and Joint Infection 100</p><p>Overview of Animal Models 103</p><p>Direct Inoculation with Minimal Trauma 103</p><p>Animal Models of Bone and Joint Infection Incorporating Trauma 108</p><p>Hematogenous Models 108</p><p>Future Directions 109</p><p>Conclusions 110</p><p>Key Points 110</p><p>References 110</p><p>Contents vii</p><p>Chapter 8 Native Joint Arthritis in Children 117</p><p>Pablo Yagupsky</p><p>Introduction 117</p><p>Epidemiology 117</p><p>Microbiology 118</p><p>Pathogenesis 122</p><p>Clinical Presentation 123</p><p>Laboratory Investigation 125</p><p>Imaging Studies 127</p><p>Differential Diagnosis 128</p><p>Treatment 128</p><p>Prognosis 133</p><p>Key Points 134</p><p>References 134</p><p>Chapter</p><p>diagnostic exchange of mobile parts with consecutive sonication fluid culture may</p><p>be indicated. Nevertheless, every intervention represents a risk for superinfection, and</p><p>should therefore be avoided. If the pain is under adequate control with analgesics, and the</p><p>patient does not urge for immediate intervention, a repeated arthrocentesis after one to</p><p>three months should be considered.</p><p>Intraoperative diagnostics including periprosthetic tissue histopathology and microbi-</p><p>ology as well as sonication fluid culture of the removed implants are more sensitive than</p><p>preoperative arthrocentesis. Therefore, these procedures are nowadays considered the gold</p><p>standard in the diagnosis of PJI [19,30,32]. Details regarding these diagnostic methods are</p><p>discussed in Chapter 11. In delayed and late PJI and in polymicrobial infections, as well as</p><p>in PJI caused by low‐virulent pathogens, the detection rate is significantly higher in intra-</p><p>operatively collected specimens compared to synovial fluid culture [33].</p><p>Key Points</p><p>● The diagnostic workup of bone and joint infections aims at identifying the pathogen-</p><p>esis and acuity of the infection, as well as the responsible pathogen, in order to plan</p><p>an adequate treatment strategy.</p><p>● Most common pathogens causing bone and joint infections originate from the</p><p>patient’s skin and mucosal microbiome; their proportion vary depending on the route</p><p>of infection.</p><p>● Accurate intraoperative tests are histopathological and microbiological examination</p><p>including synovial fluid, periprosthetic tissue, and sonication fluid culture.</p><p>● In case of septic patients with bone and joint infections, blood cultures should be col-</p><p>lected before initiation of antimicrobial treatment in order to diagnose primary or, less</p><p>frequently, secondary bacteremia. In case of suspected hematogenous periprosthetic</p><p>joint infection, the isolated pathogen guides the rational further diagnostic workup to</p><p>detect the primary focus.</p><p>References</p><p>1. Zeller V, Kerroumi Y, Meyssonnier V, et al. Analysis of postoperative and hematogenous pros-</p><p>thetic joint‐infection microbiological patterns in a large cohort. J Infect. 2018;76(4):</p><p>328–334.</p><p>2. Rakow A, Perka C, Trampuz A, et al. Origin and characteristics of haematogenous peripros-</p><p>thetic joint infection. Clin Microbiol Infect. 2019;25(7):845–850.</p><p>3. Tande AJ, Patel R. Prosthetic joint infection. Clin Microbiol Rev. 2014;27(2):302–345.</p><p>18 Bone and Joint Infections</p><p>4. Corvec S, Portillo ME, Pasticci BM, et al. Epidemiology and new developments in the diagno-</p><p>sis of prosthetic joint infection. Int J Artif Organs. 2012;35(10):923–934.</p><p>5. Koder K, Hardt S, Gellert MS, et al. Outcome of spinal implant‐associated infections treated</p><p>with or without biofilm‐active antibiotics: results from a 10‐year cohort study. Infection. 2020.</p><p>(Epub ahead of print)</p><p>6. Margaryan D, Renz N, Bervar M, et al. Spinal implant‐associated infections: A prospective</p><p>multicenter cohort study. Int J Antimicrob Agents. 2020:106116.</p><p>7. Widmer AF, Frei R, Rajacic Z, et al. Correlation between in vivo and in vitro efficacy of anti-</p><p>microbial agents against foreign body infections. J Infect Dis. 1990;162(1):96–102.</p><p>8. Zimmerli W. Experimental models in the investigation of device‐related infections. J Antimicrob</p><p>Chemother. 1993;31 Suppl D:97–102.</p><p>9. Lowik CAM, Zijlstra WP, Knobben BAS, et al. Obese patients have higher rates of polymicro-</p><p>bial and Gram‐negative early periprosthetic joint infections of the hip than non‐obese patients.</p><p>PLoS One. 2019;14(4):e0215035.</p><p>10. Ohl CA, Forster D. Infectious arthritis of native joints. In: Bennett JE, Dolin R, Blaser MJ,</p><p>et al. Mandell, Douglas, and Bennett’s Principles and Practice of Infectious Diseases. 8th edi-</p><p>tion, pp 1302–1317, ed. Philadelphia, PA: Elsevier/Saunders; 2015.</p><p>11. Mylona E, Samarkos M, Kakalou E, et  al. Pyogenic vertebral osteomyelitis: a systematic</p><p>review of clinical characteristics. Semin Arthritis Rheum. 2009;39(1):10–17.</p><p>12. Yusuf E, Steinrucken J, Buchegger T, et al. A descriptive study on the surgery and the microbi-</p><p>ology of Gustilo type III fractures in an university hospital in Switzerland. Acta Orthop Belg.</p><p>2015;81(2):327–332.</p><p>13. Giesecke MT, Schwabe P, Wichlas F, et al. Impact of high prevalence of pseudomonas and</p><p>polymicrobial Gram‐negative infections in major sub‐/total traumatic amputations on empiric</p><p>antimicrobial therapy: a retrospective study. World J Emerg Surg. 2014;9(1):55.</p><p>14. Zimmerli W. Clinical practice. Vertebral osteomyelitis. N Engl J Med. 2010;362(11):</p><p>1022–1029.</p><p>15. Sampedro MF, Huddleston PM, Piper KE, et al. A biofilm approach to detect bacteria on</p><p>removed spinal implants. Spine. 2010;35(12):1218–1224.</p><p>16. Prinz V, Bayerl S, Renz N, et al. High frequency of low‐virulent microorganisms detected by</p><p>sonication of pedicle screws: a potential cause for implant failure. J Neurosurg Spine.</p><p>2019;31(3):424–429.</p><p>17. Rutges JPHJ, Kempen DH, van Dijk M, et al. Outcome of conservative and surgical treatment</p><p>of pyogenic spondylodiscitis: a systematic literature review. Eur Spine J. 2015;</p><p>25(4):983–999.</p><p>18. Berbari EF, Kanj SS, Kowalski TJ, et al. Executive Summary: 2015 Infectious Diseases Society</p><p>of America (IDSA) Clinical Practice Guidelines for the Diagnosis and Treatment of Native</p><p>Vertebral Osteomyelitis in Adults. Clin Infect Dis. 2015;61(6):859–863.</p><p>19. Onsea J, Depypere M, Govaert G, et al. Accuracy of tissue and sonication fluid sampling for</p><p>the diagnosis of fracture‐related infection: A Systematic Review and Critical Appraisal. J Bone</p><p>Jt Infect. 2018;3(4):173–181.</p><p>20. Metsemakers WJ, Morgenstern M, McNally MA, et al. Fracture‐related infection: A consensus</p><p>on definition from an international expert group. Injury. 2018;49(3):505–510.</p><p>21. Govaert GAM, Kuehl R, Atkins BL, et al. Diagnosing fracture‐related infection: Current con-</p><p>cepts and recommendations. J Orthop Trauma. 2020;34(1):8–17.</p><p>22. Depypere M, Morgenstern M, Kuehl R, et  al. Pathogenesis and management of fracture‐</p><p>related infection. Clin Microbiol Infect. 2020;26(5):572–578.</p><p>23. Morgenstern M, Athanasou NA, Ferguson JY, et al. The value of quantitative histology in the</p><p>diagnosis of fracture‐related infection. Bone Joint J. 2018;100‐b(7):966–972.</p><p>24. Conen A, Borens O. Septic Arthritis. In: Kates SL, Borens O, editors. Principles of Orthopedic</p><p>Infection Management. Davos, Switzerland: AO Publishing; 2016. p. 213–226.</p><p>25. Shah K, Spear J, Nathanson LA, et al. Does the presence of crystal arthritis rule out septic</p><p>arthritis? J Emerg Med. 2007;32(1):23–26.</p><p>2 Diagnostic Approach in Bone and Joint Infections 19</p><p>26. Baillet A, Trocme C, Romand X, et al. Calprotectin discriminates septic arthritis from pseud-</p><p>ogout and rheumatoid arthritis. Rheumatology (Oxford, England). 2019;58(9):1644–1648.</p><p>27. Gratacos J, Vila J, Moya F, et al. D‐lactic acid in synovial fluid. A rapid diagnostic test for</p><p>bacterial synovitis. J Rheumatol. 1995;22(8):1504–1508.</p><p>28. Lenski M, Scherer MA. Analysis of synovial inflammatory markers to differ infectious from</p><p>gouty arthritis. Clin Biochem. 2014;47(1–2):49–55.</p><p>29. Mouzopoulos G, Fotopoulos VC, Tzurbakis M. Septic knee arthritis following ACL reconstruc-</p><p>tion: a systematic review. Knee Surg Sports Traumatol Arthrosc. 2009;17(9):1033–1042.</p><p>30. Dudareva M, Barrett L, Figtree M, et al. Sonication versus tissue sampling for diagnosis of</p><p>prosthetic joint and other orthopedic device‐related infections. J Clin Microbiol. 2018;56(12).</p><p>31. Renz N, Yermak K., Perka C., Trampuz A. Alpha defensin lateral flow test for diagnosis of</p><p>periprosthetic joint infection. Not a screening but a confirmatory test. J Bone Joint Surg Am.</p><p>2018(100(9)):742–750.</p><p>32. Krenn V, Morawietz L, Perino G, et al. Revised histopathological consensus classification of</p><p>joint implant related pathology. Pathol Res Pract. 2014;210(12):779–786.</p><p>33. Schulz P, Dlaska CE, Perka C, et al. Preoperative synovial fluid culture poorly predicts the</p><p>pathogen causing periprosthetic joint infection. Infection.</p><p>2020 Nov 3. doi: 10.1007/s15010-</p><p>020-01540-2. Epub ahead of print.</p><p>21</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>Chapter 3</p><p>Introduction</p><p>The association of certain microorganisms such as Staphylococcus epidermidis,</p><p>Staphylococcus aureus, and ß‐hemolytic streptococci with periprosthetic joint infection</p><p>(PJI) has been recognized for many years. These microorganisms are usually easily</p><p>isolated on culture using conventional media.</p><p>Other microorganisms that can cause PJI may be more difficult to culture or are not</p><p>typically associated with PJI. Recent advances in molecular microbiology and culture</p><p>procedures have led to the discovery of microorganisms that are less commonly associ-</p><p>ated with PJI. The recognition of PJI due to these microorganisms is important because</p><p>the course of untreated PJI is consistently unfavorable.</p><p>Gram‐Positive Microorganisms</p><p>Staphylococcus lugdunensis</p><p>S. lugdunensis is a coagulase‐negative staphylococcus (CNS) whose pathogenic potential is</p><p>similar to that of S. aureus. It also develops a biofilm phenotype which may appear as small‐</p><p>colony variant [1]. Although genetically indistinguishable, small‐colony variants differ in size</p><p>and antibiotic susceptibility from the parent strain, and are responsible for chronic persistent</p><p>infections and failure of antibiotic treatment. Several cases of PJIs after total knee (TKA)</p><p>and total hip arthroplasty (THA) due to S. lugdunensis have been reported, in both immu-</p><p>nocompromised (rheumatoid arthritis, multiple myeloma, pancreatic cancer, diabetes melli-</p><p>tus) and immunocompetent patients. The microorganism more often causes PJI after TKA</p><p>than THA. Presentation of PJIs due to Staphylococcus lugdunensis can be acute, with fever</p><p>and local signs of inflammation or persistent pain at the surgical site.</p><p>Unusual Microorganisms</p><p>in Periprosthetic Joint Infection</p><p>Camelia Marculescu and Werner Zimmerli</p><p>22 Bone and Joint Infections</p><p>Microbiologic Diagnosis</p><p>S. lugdunensis frequently produces a clumping factor that may result in positive slide</p><p>(short) coagulase test. Therefore, identification of S. lugdunensis should include tube</p><p>(long) coagulase test. Additional biochemical tests (positive results on pyrrolidonyl‐</p><p>arylamidase and ornithine‐decarboxylase tests) are helpful for final identification.</p><p>Therapy</p><p>Therapy for S. lugdunensis PJI included two‐stage exchange (TSE), debridement and</p><p>retention of implant (DAIR), and one‐stage exchange (OSE). There are insufficient data</p><p>to determine whether the outcome of S. lugdunensis PJI treated with DAIR will be similar</p><p>to the outcome of S. aureus PJI [1–4].</p><p>Streptococcus bovis</p><p>The association between S. bovis bacteremia and the presence of an underlying colonic</p><p>pathology (colorectal carcinoma, polyps, colonic ulcers) has been extensively reported [5,6].</p><p>Some rare cases of S. bovis PJIs were reported in the literature [7,8]. Emerton et al. [8]</p><p>reported one case of S. bovis THA PJI occurring one month after a nonspecific flu‐like</p><p>illness, in the absence of endocarditis.</p><p>Microbiologic Diagnosis</p><p>Colonies of S. bovis are typically non‐hemolytic on sheep blood agar. S. bovis grows in</p><p>40% bile and hydrolyzes esculin. It can be distinguished from enterococci because the</p><p>former is PYR negative, fails to hydrolyze arginine, and is unable to grow in 6.5% salt</p><p>broth. S. bovis can be further divided into biotype I and II, based on additional biochemical</p><p>reactions. 16S rRNA sequencing allows further differentiation of biotype II strains into</p><p>groups 1 and 2 [9]. The microorganism is usually susceptible to penicillin.</p><p>Therapy</p><p>Antimicrobial therapy without revision surgery was unsuccessful in two cases with</p><p>S.  bovis PJI, as could be expected. Subsequently, both patients underwent TSE with</p><p>favorable outcome after 74 and 6 months of follow‐up, respectively [7,8]. In the series of</p><p>Thompson et al. [6], patients with acute PJI were treated with DAIR, and those with</p><p>chronic PJI with TSE. Both patients treated with DAIR and two patients with TSE got</p><p>suppressive therapy for unknown reasons. In the five episodes managed with TSE, treat-</p><p>ment was successful. A six‐week course of ceftriaxone was found to be effective [6].</p><p>A  majority of the reported cases had colonic pathology. Thus, colonic investigation</p><p>should be pursued for patients diagnosed with S. bovis PJI.</p><p>Gemella</p><p>G. morbillorum (formerly Streptococcus morbillorum), G. haemolysans, and G. sanguinis</p><p>PJI have been reported [10–12]. G. morbillorum is rarely involved in human infections</p><p>such as endocarditis. G. sanguinis was reported in one case as a late acute infection [12].</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 23</p><p>Microbiologic Diagnosis</p><p>G. morbillorum is an anaerobic to aerotolerant Gram‐positive coccus. Specimens</p><p>collected at the time of surgery should be inoculated on aerobe and anaerobe plates.</p><p>Improving the rate of isolation of such fastidious microorganisms from the joint fluid</p><p>may require direct inoculation of joint fluid in blood culture bottles [12,13]. G. morbil-</p><p>lorum is identified as a Gram‐positive, pleomorphic, slow‐growing, and initially anaero-</p><p>bic bacteria. Identification of G. morbillorum may be impeded by its slow growth rate, as</p><p>well as variable morphology and staining properties. Identification of G. haemolysans is</p><p>usually established by a series of biochemical reactions (API 20 Strep identification sys-</p><p>tem) together with Gram stain morphology. Partial 16s rRNA sequencing identified</p><p>G. sanguinis [12]. Better aerobic growth allows differentiation of G. haemolysans from</p><p>viridans group streptococci and G. morbillorum. Gemellae are typically susceptible to</p><p>penicillin, erythromycin, tetracycline, and vancomycin.</p><p>Therapy</p><p>Reported cases were treated with TSE or DAIR. One patient with G. haemolysans TKA</p><p>PJI was treated with resection arthroplasty, six weeks of parenteral penicillin G, followed</p><p>by reimplantation with favorable outcome after two months of follow‐up [10]. Chronic</p><p>oral suppression with doxycycline was used in the case treated with DAIR [12].</p><p>Listeria monocytogenes</p><p>L. monocytogenes is a Gram‐positive, nonspore‐forming aerobic rod that causes a variety</p><p>of infections in humans. PJI due to Listeria is very rare. The source of listeriosis is usually</p><p>unknown, but a number of reports have implicated the consumption of unpasteurized</p><p>milk/cheese, vegetables, or processed meat. L. monocytogenes PJI tend to occur in elderly</p><p>patients or immunocompromised patients with malignancy, transplantation, diabetes,</p><p>cirrhosis or rheumatoid arthritis. PJI caused by Listeria spp. typically presents as late</p><p>hematogenous infection.</p><p>Microbiologic Diagnosis</p><p>Diagnosis is based on Gram stain and culture of the microorganism. Listeria produces a</p><p>characteristic appearance on sheep blood agar with small zones of clear beta hemolysis</p><p>around each colony. In addition, saline suspensions of Listeria grown in vitro demon-</p><p>strate characteristic tumbling motility, whereas Corynebacterium spp. (ie. diphtheroids)</p><p>do not exhibit motility. Listeria grows well at refrigeration temperatures (4° to 10°C).</p><p>Blood cultures can be negative in cases of PJI [14, 15]. In vitro, Listeria is susceptible</p><p>to ampicillin, trimethoprim‐sulfamethoxazole, aminoglycosides, and vancomycin.</p><p>Cephalosporins are not effective against Listeria and false positive in vitro sensitivity</p><p>reports can result from the use of disk diffusion tests [15,16].</p><p>Therapy</p><p>Prosthesis removal was the surgical modality that eventually led to cure of the infection</p><p>in five cases. DAIR and OSE followed by antimicrobial suppression were also reported</p><p>24 Bone and Joint Infections</p><p>with good short‐term outcomes [15,17,18]. Failure of conservative management requiring</p><p>OSE was recently reported after eight months of follow‐up. This patient had no signs of</p><p>infection after</p><p>one year of follow‐up [19]. The recommended medical therapy is parenteral</p><p>ampicillin with or without gentamicin. Gentamicin should not be used as monotherapy</p><p>[16,17]. Little information about the optimal duration of treatment is available. Six weeks</p><p>of intravenous antimicrobials appear to be adequate in reported cases. Relapses can</p><p>occur because of the involvement of foreign material and the intracellular growth of the</p><p>microorganism. The duration of chronic oral antimicrobial suppression in cases treated</p><p>with prosthesis retention needs to be assessed carefully, especially in immunocompro-</p><p>mised patients. Profoundly immunocompromised patients may require life‐long oral</p><p>antimicrobial suppression.</p><p>Nocardia</p><p>Nocardia ssp. are opportunistic microorganisms, causing infection mainly in immuno-</p><p>compromised patients (cancer, hematological malignancies, particularly lymphoid,</p><p>chronic respiratory diseases, corticosteroid treatment, transplant recipients, sarcoidosis).</p><p>Several cases of Nocardia (nova, asteroids, and cyriacigeorgica) THA and TKA PJIs have</p><p>been reported, two in immunocompetent patients, and the remainder in immunocompro-</p><p>mised patients [20–23]. Nocardia nova presented as a late, indolent infection in a patient</p><p>with systemic lupus erythematosus, two years after being diagnosed with pulmonary</p><p>nocardiosis. In an immunocompetent patient, Nocardia PJI was presumably caused peri-</p><p>operative contamination [21].</p><p>Microbiologic Diagnosis</p><p>Diagnosis is based on the typical morphology on the modified acid‐fast stain. It appears</p><p>as a thin, Gram‐positive microorganism with branching filaments with a beaded appear-</p><p>ance. The acid‐fast reaction and the production of aerial branching allow differentiation</p><p>of Nocardia from other aerobic and anaerobic actinomycetes. When deep surgical cul-</p><p>tures are obtained, they are usually positive 85–90% of the time [21,24]. However, even</p><p>when adequate specimens are obtained, recovery of Nocardia in the laboratory can be</p><p>difficult. Most routine bacterial, fungal, or mycobacterial culture media can support the</p><p>growth of Nocardia. However, Nocardia requires prolonged incubation of up to two to</p><p>three weeks. Speciation of the different taxons of Nocardia is difficult. The optimal</p><p>method for antimicrobial susceptibility testing has not been established, and there are no</p><p>specific NCCLS MIC breakpoints for Nocardia [21,25]. It is therefore recommended to</p><p>send the Nocardia isolates to a reference laboratory. In the reported cases of PJI, N. nova</p><p>was susceptible to imipenem, erythromycin, and amikacin. In contrast to N. asteroides,</p><p>N. nova is resistant to trimethoprim‐sulfamethoxazole. Although N. nova may demon-</p><p>strate susceptibility to ampicillin or amoxicillin in vitro, it consistently carries a beta‐</p><p>lactamase, which may hydrolyze these antibiotics [21,25].</p><p>Therapy</p><p>The drugs of choice for the treatment of N. asteroides infections are either sulfonamides or</p><p>tetracyclines. The suggested duration of therapy to prevent relapses for cutaneous and</p><p>pulmonary disease is three months, particularly with trimethoprim‐sulfamethoxazole</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 25</p><p>combinations [21,26]. However, the efficacy of combinations remains controversial,</p><p>particularly against N. nova. Imipenem‐amikacin combination appears to be more effective</p><p>than trimethoprim‐sulfamethoxazole in nocardiosis, in particular N. nova infection.</p><p>Linezolid is active in vitro against all Nocardia spp. TSE and even OSE followed by chronic</p><p>oral suppression have yielded favorable outcomes in the reported cases [20–23].</p><p>Dietzia maris</p><p>D. maris is an environmental actinomycete that is rarely involved in human disease. To</p><p>date, only one case of THA infection has been reported. D. maris was presumably</p><p>acquired as superinfection during a prolonged free interval with a spacer [27].</p><p>Microbiologic Diagnosis</p><p>Diagnosis is based on the Gram‐stain from intraoperative specimens that disclosed many</p><p>Gram‐positive cocci germinating into short rods. The isolate was identified as D. maris</p><p>using the API Coryne strip. Rapid molecular methods and cellular fatty acid analysis</p><p>confirmed the identification of D. maris. The microorganism was susceptible, by disk‐</p><p>diffusion method, to amoxicillin, imipenem, gentamicin, trimethoprim‐sulfamethoxa-</p><p>zole, rifampin, clindamycin, vancomycin, and pristinamycin.</p><p>Therapy</p><p>This patient was treated with teicoplanin for four months without further surgery. No</p><p>follow‐up data were reported.</p><p>Tsukamurella</p><p>T. paurometabolum is a Gram‐positive, weakly or variably acid‐fast, nonmotile, obligate</p><p>aerobic bacillus that exists primarily as a saprophyte in the soil. Tsukamurella should be</p><p>recognized as a potential pathogen in patients with immunosuppression, indwelling for-</p><p>eign bodies, and postoperative wounds. In one case, Tsukamurella was found after</p><p>repeated debridements and removal of a TKA for a mixed (Peptostreptococcus and CNS)</p><p>infection [28].</p><p>Microbiologic Diagnosis</p><p>Diagnosis was based on bone culture results that yielded Gram‐positive rods after 13</p><p>days of incubation. Final identification was done five weeks later. The microorganism is</p><p>susceptible to sulfamethoxazole, clarithromycin, imipenem, amikacin, ciprofloxacin,</p><p>rifampin, vancomycin, and third‐generation cephalosporins.</p><p>Therapy</p><p>The reported case was treated with TSE and a two‐month course of clarithromycin plus</p><p>ciprofloxacin. Reimplantation was performed four months after explantation. The out-</p><p>come was successful, however, without mentioning the duration of follow‐up.</p><p>26 Bone and Joint Infections</p><p>Oerskovia</p><p>Oerskovia are Gram‐positive, Nocardia‐like bacilli which inhabit the soil and rarely cause</p><p>human infections. One case of late, presumably exogenous TKA‐PJI with O. xanthineo-</p><p>lytica, has been reported [29]. The patient suffered from intermittent pain and swelling</p><p>after a knee trauma. Revision arthroplasty was performed for presumed aseptic loosen-</p><p>ing, but O. xanthineolytica was diagnosed at removal of the implant.</p><p>Microbiologic Diagnosis</p><p>Diagnosis was based on the Gram stain and culture of the surgical specimens, as well as</p><p>biochemical tests using API Coryne strip. Gram stain revealed Gram‐positive branching</p><p>diphtheroid rods. Susceptibility testing for many of these fastidious, slow growing micro-</p><p>organisms is not yet standardized and differences in the MIC of the microorganism with</p><p>different method used might be encountered [29]. The most active antimicrobial in vitro</p><p>is vancomycin [29,30]. Penicillin or ampicillin, rifampin, and vancomycin are drugs of</p><p>choice in infections caused by this microorganism.</p><p>Therapy</p><p>The patient was treated with TSE and a five‐week course of trimethoprim‐sulfamethoxazole.</p><p>Reimplantation was performed three months after explantation. The outcome was</p><p>excellent after six months of follow‐up.</p><p>Bacillus alvei</p><p>Infections with non‐anthrax‐Bacillus spp. generally occur in immunocompromised</p><p>patients or intravenous drug users. A 26‐year‐old woman with sickle‐cell disease suffered</p><p>from B. alvei sepsis seeding on a THA 12 years after implantation. The patient was treated</p><p>with removal of the prosthesis and a six‐week course of vancomycin. The patient</p><p>remained well after 18 months of follow‐up [31].</p><p>Bacillus cereus</p><p>B. cereus is a catalase‐positive, aerobic, spore‐forming bacillus. Soft tissue and bone</p><p>infections due to this pathogen have been associated with trauma, intravenous drug use,</p><p>and immunodeficiency. B. cereus isolated from blood has been easily dismissed as con-</p><p>taminant without repeated isolation from multiple blood cultures. A THA PJI due to</p><p>Bacillus cereus and bacteremia occurring 13 years after THA has been reported in a dia-</p><p>betic female [32]. The patient was treated with implant removal followed by intravenous</p><p>vancomycin. The outcome cannot be judged due to the very short follow‐up.</p><p>Gram‐Negative Microorganisms</p><p>Salmonella</p><p>Nontyphoidal Salmonella infection is mainly seen in young or debilitated</p><p>patients, in</p><p>patients with sickle cell disease, collagen vascular disease, immunosuppressive medica-</p><p>tions, or HIV. Among reported cases with PJI, mainly after THA, one patient had sickle</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 27</p><p>cell trait and others had different immunocompromising conditions such as rheumatoid</p><p>arthritis, renal transplantation, familial Mediterranean fever, ankylosing spondylitis,</p><p>immunosuppressive meds, chronic lymphatic leukemia, ulcerative colitis, lung adenocarci-</p><p>noma, and polymyositis [33–40]. S. typhimurium was the most frequently isolate seen.</p><p>Other reported Salmonella species to cause PJI were S. dublin, newport, muenchen, hirschfeldii,</p><p>enteritidis, and choleraesuis. Salmonella PJI is typically hematogenously seeded from the</p><p>gastrointestinal tract. The presentation is acute, usually secondary to bacteremia and/or</p><p>gastroenteritis. Infections may either occur in the early or late postoperative period.</p><p>Microbiologic Diagnosis</p><p>Salmonella are relatively easy to identify in the clinical microbiology laboratory. They</p><p>grow under both aerobic and anaerobic conditions. Salmonella are oxidase and lactose‐</p><p>negative (clear to semitranslucent colonies on MacConkey agar plates). S. typhi, S. para-</p><p>typhi C, and S. dublin strains have the Vi antigen. A DNA sequence encoding the Vi</p><p>antigen was used in developing a nested PCR for S. typhi [41]. In one study it was noted</p><p>that the prevalence of S. typhimurium isolates with resistance to ampicillin, chloram-</p><p>phenicol, streptomycin, sulfonamides, and tetracycline was 34% in 1996 [42]. In 2001,</p><p>60% of S. cholearaesuis isolates were resistant to ciprofloxacin [43]. Given the increasing</p><p>rate of resistance among Salmonella strains, therapy should be guided by the in vitro</p><p>susceptibility studies. For susceptible isolates, fluoroquinolones may be the preferred</p><p>agent, since they have activity against biofilm bacteria [46].</p><p>Therapy</p><p>The preferred surgical therapy is removal of prosthesis followed by antimicrobial therapy.</p><p>Day et al. [44] reported a case of a TKA infection due to Salmonella enteritidis that was</p><p>treated with DAIR and exchange of mobile parts, followed by six weeks of ceftriaxone.</p><p>Chronic oral antimicrobial suppression was not used. In this case the outcome was excel-</p><p>lent after six years of follow‐up. In another case, a patient with S. dublin THA infection</p><p>was also treated with DAIR. Trimethoprim‐sulfamethoxazole was administered for two</p><p>years. The patient did not show any clinical signs of recurrence [45]. One case of S. dublin</p><p>THA PJI relapsed after OSE. Ciprofloxacin was administered for one year, and the infec-</p><p>tion was cured after three and a half years of follow‐up [46]. TSE for Salmonella THA</p><p>infection was associated with an excellent outcome in two patients [47,48]. In three cases</p><p>no surgical intervention was done, and patients were maintained on chronic oral suppres-</p><p>sion [49–51]. The infection relapsed in one patient. As could be expected, antimicrobial</p><p>therapy alone is insufficient to cure PJI.</p><p>Neisseria meningitidis</p><p>Primary meningococcal arthritis is an unusual form of disseminated meningococcal dis-</p><p>ease in which the features of acute pyogenic arthritis develop without meningitis or</p><p>meningococcemia. Three cases of PJI due to Neisseria meningitides were reported in the</p><p>literature. All three involved a TKA and were treated with DAIR [52,53,54]. Vikram et al.</p><p>[52] described an 80‐year‐old woman without obvious risk factors for PJI or meningococ-</p><p>cal disease. She suffered from primary meningococcal TKA PJI. The onset of symptoms</p><p>was acute, and the patient had associated meningococcal bacteremia, without evidence</p><p>of meningitis.</p><p>28 Bone and Joint Infections</p><p>Microbiologic Diagnosis</p><p>Neisseria meningitidis will grow both on sheep blood agar and chocolate agar or choco-</p><p>late agar containing antibiotics (e.g. Thayer‐Martin agar). N. gonorrhoeae does not grow</p><p>on sheep blood agar. Both N meningitidis and N. gonorrhoeae utilize glucose. N. menin-</p><p>gitidis is differentiated from N. gonorrhoeae by utilization of maltose.</p><p>Therapy</p><p>All patients were treated with DAIR, followed by variable courses of parenteral antimicrobial</p><p>therapy (ranging 3–6 weeks), followed by a short course of oral amoxicillin or ciprofloxacin</p><p>for six to nine weeks [53,54], or indefinite chronic oral suppression with penicillin [52].</p><p>Haemophilus spp.</p><p>There are sporadic reported cases of Haemophilus influenzae THA and TKA PJIs [55,56].</p><p>Risk factors for Haemophilus influenzae arthritis include multiple myeloma, systemic</p><p>lupus erythematous, rheumatoid arthritis, chronic lymphatic leukemia, common variable</p><p>hypogammaglobulinemia, diabetes mellitus, and alcohol abuse. A depressed immune</p><p>status may play a role in the development of such infections. Blood cultures are usually</p><p>positive. All reported cases were acute hematogenous infections.</p><p>Microbiologic Diagnosis</p><p>Diagnosis is based on the Gram stain from a joint aspirate showing Gram‐negative coc-</p><p>cobacilli. H. influenzae grows aerobically as pinpoint colonies on chocolate blood agar.</p><p>Speciation is done by the differential growth requirements for hematin (X factor) and</p><p>nicotinamide factor (V factor).</p><p>Therapy</p><p>Patients were treated with DAIR, since they had no evidence of prosthesis loosening and</p><p>the infections were acute. Patients received intravenous antimicrobial therapy. An oral</p><p>fluoroquinolone was used for suppression during one year in one case [57]. One patient</p><p>required resection arthroplasty, followed by six weeks of IV‐ampicillin and long‐term</p><p>oral amoxicillin [58]. The outcome of H. influenzae PJI was favorable, after a follow‐up</p><p>between six months and five years. Routine vaccination against Haemophilus influenza</p><p>serotype B may be helpful in patients with prosthetic joints and hematological malignan-</p><p>cies predisposing to bacterial infections [56].</p><p>A patient with two prosthetic hip joints suffered from bilateral PJI caused by</p><p>Haemophilus parainfluenza after extraction of several loosening teeth without prophy-</p><p>laxis [59].</p><p>Microbiologic diagnosis relies on Gram‐stain, aerobic culture of the microorganism</p><p>on chocolate agar and determination of X and V factor requirements for growth. H.</p><p>parainfluenzae is usually susceptible to ampicillin, cefuroxime, ceftriaxone, quinolones</p><p>and trimethoprim‐sulfamethoxazole.</p><p>Septic loosening occurred after three months in one case, and required TSE for</p><p>cure [59]. Another patient with TKA PJI due to H. parainfluenzae after dental surgery</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 29</p><p>improved after an eight‐week course of intravenous ceftriaxone, followed by oral</p><p>antimicrobials (trimethoprim‐sulfamethoxazole and subsequently oral ciprofloxacin)</p><p>for two years [60].</p><p>Zoonotic Microorganisms</p><p>Brucella spp.</p><p>PJI due to Brucella is infrequent. Brucella melitensis infection in humans is transmitted</p><p>from animals through contaminated dairy products, nonpasteurized milk, by contact</p><p>through a skin laceration, or by inhalation. To date, 35 cases of Brucella PJI have been</p><p>reported [61–66]. Almost all patients consumed unpasteurized dairy products or had</p><p>occupational exposure (farmers, contact with cattle or goats, wild meat butchers).</p><p>Brucella PJI presents as an indolent infection, mainly with local symptoms. Systemic</p><p>symptoms were present in only 13/34 (38%). One patient had a sinus tract [64]. The</p><p>median time from prosthesis implantation to diagnosis of PJI was 48 months (range</p><p>between 0 and 168 months). More than half of the patients had documented loosening</p><p>on radiographs. Brucella melitensis was isolated in the vast majority of patients (25/35),</p><p>followed by Brucella abortus (4/35), suis (1 case), and Brucella spp. (5/35).</p><p>Microbiologic Diagnosis</p><p>Diagnosis was mostly made by positive cultures in joint fluid, rarely by blood cultures, or</p><p>intraoperative tissue biopsies. Coinfection with other pathogens is reported. A negative joint</p><p>culture result cannot rule out osteoarticular</p><p>brucellosis. Using the BACTEC blood culture</p><p>system (Becton Dickinson), detection of brucellae can be accomplished with a sensitivity of</p><p>95% within 6 days [67]. Blood PCR‐ELISA for Brucella is more sensitive (94.9%) and spe-</p><p>cific (96.5%) than conventional manual blood culture systems. Diagnosis of Brucella spp.</p><p>infection is based on identification of Gram‐negative, oxidase‐positive cocci and rods. Once</p><p>suspicious colonies are isolated, differentiation of Brucella species is usually accomplished</p><p>by agglutination with specific antiserum. An agglutination test of the serum and/or the</p><p>synovial fluid to determine Brucella antibodies with titer greater than 1:160 is considered</p><p>evidence for the presence of Brucella infection [65]. Synovial α‐defensin test was reported</p><p>negative in a case of documented bacteremic Brucella melitensis bilateral TKA PJI [68].</p><p>Therapy</p><p>Management of Brucella osteomyelitis and PJI is controversial with regard to antimicro-</p><p>bial selection, duration of therapy, and the role of surgery. A recent meta‐analysis showed</p><p>that combination therapy with doxycycline/aminoglycoside (streptomycin or gentamicin)</p><p>is preferred over doxycycline/rifampin or other combinations [69]. The optimal treatment</p><p>duration is unknown, ranging from 6 weeks to 19 months in the current series. Malizos</p><p>and al. [65] suggest longer duration of therapy of patients with joint implants, even if the</p><p>Brucella blood titers become negative. They suggest that treatment efficacy should be</p><p>monitored through serum and joint‐aspirate Brucella titers.</p><p>Monotherapy with a fluoroquinolone or trimethoprim‐sulfamethoxazole has an unac-</p><p>ceptable high‐risk rate of relapse [70]. Complete eradication of the microorganism is</p><p>30 Bone and Joint Infections</p><p>difficult to achieve, and relapse does occur, especially when the disease is caused by</p><p>B. melitensis. The most frequent cause of relapse includes failure to complete treatment</p><p>and unrecognized localized foci of infection [66]. The relapse is confirmed by the isola-</p><p>tion of brucellae from blood or other tissues of a patient with recurrent symptoms.</p><p>TSE was performed in 13/35 cases. The optimal time to reimplantation is unknown,</p><p>because there is no consistent test to ensure successful eradication. Generally, patients</p><p>with loose prostheses were treated with a TSE with a long interval (6 weeks–6 months</p><p>between the stages). Prolonged antimicrobial therapy without resection of the prosthesis</p><p>was chosen in cases without implant loosening (9/35), and yielded some success. DAIR</p><p>and OSE or resection arthroplasty alone were also reported. Treatment was administered</p><p>until resolution of infection and sterilization of synovial fluid cultures. Failure of antimi-</p><p>crobial therapy was documented in one case of B. abortus THA infection [71]. The patient</p><p>was subsequently treated with OSE followed by one year of combination antimicrobial</p><p>therapy until negative serologic titers were reached. The duration of follow‐up for all</p><p>reported cases varied from six months to two years.</p><p>Francisella tularensis</p><p>PJI caused by F. tularensis was recently described [72–74]. Cooper et al. [73] reported a case</p><p>of a chronic TKA PJI in a 68‐year‐old man with rheumatoid arthritis treated with metho-</p><p>trexate who had a history of wood tick bite. He presented six months after TKA with per-</p><p>sistent, low‐grade infection. It was presumed that asymptomatic lympho‐hematogenous</p><p>seeding of F. tularensis was the likely mechanism of infection. Another two cases were</p><p>reported by Chrdle et al. [72] in immunocompetent individuals in the late postoperative</p><p>period diagnosed initially with culture‐negative PJIs. Their risk factors for tularemia were</p><p>exposure to a rabbit barn and outdoor gardening in a tularemic area, respectively. A fourth</p><p>case involving a THA was recorded in a hunter with no known recent exposures to tularemia</p><p>[74]. Two of the reported cases had a lesion in the lower extremity, suggestive of a tick bite.</p><p>Microbiologic Diagnosis</p><p>F. tularensis is suspected on the basis of growth, morphology, and a weak catalase test.</p><p>The microorganism is fastidious. Most strains require cysteine or cystine for growth.</p><p>Prolonged incubation time to 14 days was recommended for culture negative PJIs [72].</p><p>Species identification can be accomplished by antisera, biochemical testing, and 16S</p><p>rRNA sequencing [75]. Antimicrobial susceptibilities have not changed in years, and</p><p>doxycycline, aminoglycosides, and fluoroquinolones alone or in combination therapy are</p><p>the mainstay of therapy. Rifampin has a relatively low MIC against F. tularensis, but</p><p>clinical experience is lacking.</p><p>Therapy</p><p>Antibiotic treatment with either ciprofloxacin/rifampin, doxycycline/gentamicin, or</p><p>doxycycline alone was given between 3 weeks and 12 months. TSE, aspiration alone with</p><p>antibiotics and DAIR without chronic oral suppression were surgical modalities</p><p>performed in the reported cases. Repeated courses of antibiotics were needed in some of</p><p>the reported cases.</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 31</p><p>Yersinia enterocolitica</p><p>To date only four cases of PJI due to Y. enterocolitica have been reported in the literature</p><p>[76,77]. This infection is primarily associated with gastrointestinal symptoms (diarrhea,</p><p>abdominal pain and mesenteric lymphadenitis in children, and fever). Severe symptoms</p><p>such as acute ileitis, myocarditis, and septicemia are more commonly reported in adults.</p><p>It may also cause reactive arthritis. Pigs and cattle are the primary reservoir of Y. entero-</p><p>colitica. Reported cases occurred in elderly patients and involved a TKA or THA several</p><p>years after implantation. The presentation of infection was rather acute with high fever</p><p>and an inflamed prosthetic joint. Only one patient with diabetes mellitus who took iron</p><p>supplements for anemia had diarrhea prior to the onset of symptoms. In this case, it is</p><p>likely that the microorganism seeded to the TKA following a transient bacteremia after</p><p>the initial enteritis Yersinia enterocolitica bacteremia without apparent source was docu-</p><p>mented in another case [77]. Iron supply associated with hemosiderin deposition due to</p><p>hemarthrosis may be the main risk factor, at least in one case [76].</p><p>Microbiologic Diagnosis</p><p>Y. enterocolitica grows well on most enteric media with the exception of Salmonella‐</p><p>Shigella agar. Serology testing by using bacterial agglutination may produce spurious</p><p>results, especially in individuals infected with serogroups other than O:8. Geographic</p><p>location and cross‐reactivity with other members of Enterobacteriacceae family, Brucella</p><p>and Rickettsia spp. are also limitations to the agglutination method. ELISA technique</p><p>has a higher sensitivity in assessing antibodies (IgG and IgA) not detected by bacterial</p><p>agglutination. However, in the absence of culture‐proven Y. enterocolitica infection, the</p><p>significance of a single positive antibody test is difficult to assess. In the reported cases,</p><p>culture from joint fluid, operative specimens, and blood, as well as serology, identified</p><p>Y. enterocolitica, serotype 0:3 and 0:9, respectively. All serogroups are susceptible to imi-</p><p>penem and aztreonam. Fluoroquinolones and broad‐spectrum cephalosporins, often in</p><p>combination with an aminoglycoside, have resulted in a successful outcome in patients</p><p>with extraintestinal Y. enterocolitica infection. The pattern of susceptibility of the four</p><p>major serogroups (O:3, O:5,27, O:8, and O:9) to ampicillin, carbenicillin, cephalothin,</p><p>cephaloridine, cephalexin, and cefoxitin was distinct to each serogroup.</p><p>Therapy</p><p>Removal of the prosthesis has been necessary for full recovery. However, a 90‐year old</p><p>patient was treated with DAIR and chronic oral suppression with ciprofloxacin, with an</p><p>initial very slow response [78]. Another case was cured with oral ciprofloxacin for six</p><p>weeks, but the duration of the follow‐up was not reported [76]. The optimal duration of</p><p>treatment is unknown.</p><p>Pasteurella multocida</p><p>P. multocida infection</p><p>typically localizes in skin and soft tissues. P. multocida, a small</p><p>Gram‐negative organism, is part of the normal mouth flora of many animals. Thirty‐two</p><p>cases of PJI due to P. multocida have been reported [79,80]. Most cases involve a TKA,</p><p>and many patients were immunocompromised patients (diabetes mellitus, rheumatoid</p><p>32 Bone and Joint Infections</p><p>arthritis, acute leukemia, breast cancer). Almost all patients presented with a history of</p><p>animal bite or animal contact or lick to the lower extremity containing the prosthetic</p><p>joint. Two TKA PJIs were described after cat bites to the upper extremity [81,82]. This</p><p>might reflect local propagation from the portal of entry via damaged lymphatic vessels or</p><p>might be secondary to hematogenous seeding [81–83]. Cases of Pasteurella PJI transmit-</p><p>ted by dog lick [84] or close contact with a dog [85] have been described. The majority of</p><p>P. multocida PJI occurred in women. The reason for gender predilection could be that</p><p>women have a higher risk of exposition as primary caretakers of pets. P. multocida PJI is</p><p>usually an acute illness.</p><p>Microbiologic Diagnosis</p><p>Pasteurella can be cultured on blood or chocolate agar, preferably with an atmosphere</p><p>containing 5% CO2. Some commercial biochemical identification systems are able to</p><p>identify strains of P. multocida. Although Pasteurella is usually susceptible to penicillin,</p><p>empiric treatment with a penicillinase‐resistant penicillin may be warranted pending in</p><p>vitro susceptibility results. This is due to the fact that some isolates are β‐lactamase pro-</p><p>ducers [86].</p><p>Therapy</p><p>Cases of P. multocida PJI were successfully treated with TSE [85,86], OSE [87,88], DAIR</p><p>[79], or resection arthroplasty [79,81].The duration of antimicrobial therapy in these</p><p>cases was variable, ranging from 3 to 10 weeks of parenteral therapy followed in some</p><p>cases by 2–3 weeks of oral antimicrobial therapy [81, 88]. Success was also reported in</p><p>cases where the infected prosthesis was retained using surgical debridement [89,90], aspi-</p><p>ration [91], or antimicrobial therapy alone [92,93]. On the other hand, several case reports</p><p>of treatment failure occurred in patients treated with antimicrobial therapy with or with-</p><p>out joint aspiration [82,84].</p><p>Guion and Sculco [86] suggest that optimal surgical therapy for these infections should</p><p>include debridement of the involved tissues and removal of all foreign material. However,</p><p>DAIR combined with prolonged antimicrobials may be sufficient for the treatment of</p><p>P multocida PJIs [79].</p><p>Some authors suggest that following cat or dog bites especially immunocompromised</p><p>patients with a total joint arthroplasty should be instructed to take a penicillin prophy-</p><p>laxis, in order to prevent hematogenous seeding on the implant [79,90].</p><p>Coxiella burnetii</p><p>Coxiella burnetii, which causes Q‐fever, is an organism that would traditionally be con-</p><p>sidered in culture‐negative PJIs. Eight cases have been described in the literature, five</p><p>involving a THA and three involving a TKA [94,95]. Only one case had a documented</p><p>contact with sheep [96]. Other cases had either no risk factors or local steroid soft‐tissue</p><p>injections or history of IVDU. Local signs were the main presenting symptoms, and one</p><p>patient was asymptomatic, Q fever being diagnosed at the time of revision THA [96]. In</p><p>six cases, the diagnosis was established by specific PCR. A serological profile with strongly</p><p>positive phase 1 IgG serology also supported the diagnosis in all cases. It is unclear</p><p>whether C burnetii is capable of biofilm formation [95]. In ostearticular infections, a</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 33</p><p>treatment of 18 months with hydroxychloroquine and doxycycline has been suggested as</p><p>the best option [97]. Alternative treatments include ciprofloxacin or trimethoprim‐</p><p>sulfamethoxazole in combination with doxycycline. Most of the PJI cases had a good</p><p>outcome. However, one case required resection arthroplasty and transfemoral amputa-</p><p>tion [98], and two others required additional surgeries [96,99].</p><p>Campylobacter</p><p>Campylobacter spp. are major causes of gastrointestinal infections in humans. Septic</p><p>native arthritis is uncommon, and usually affects elderly patients, immunocompromised</p><p>patients, and patients with alcoholism, cirrhosis, chronic lymphatic leukemia, cancer, or</p><p>diabetes mellitus. Infections caused by C. jejuni, C. coli, and C. lari can be acquired via</p><p>contact with farm animals, chickens, seagulls, and nonhuman primates. A fatal case of</p><p>C. lari THA‐PJI with bacteremia has been reported in an 81‐year‐old immunocompetent</p><p>patient who suffered from a non‐specific infectious syndrome without diarrhea for 3 weeks</p><p>[100]. Not all cases have a preceding diarrheic episode. Bacteremia is often preceded or</p><p>accompanied by diarrhea [101]. Cases of THA PJI caused by C. jejuni [102, 103], lari</p><p>[100], and fetus [104] have been reported.</p><p>Microbiologic Diagnosis</p><p>Campylobacter spp. are fastidious, slowly growing Gram‐negative bacilli which require</p><p>selective, enriched media with prolonged incubation under microaerophilic conditions.</p><p>Generally, Campylobacter spp. are not easily visualized with the safranin counterstain</p><p>commonly used in Gram staining. Therefore, its identification is difficult, especially</p><p>from a periprosthetic joint site, at which Campylobacter is infrequently or unexpectedly</p><p>isolated. Molecular testing is useful in identification of campylobacters, especially in</p><p>distinguishing C. lari from C. jejuni, since they both share common phenotypic proper-</p><p>ties [100]. A previously described C. jejuni/C. coli real time PCR assay targeting cadF</p><p>and designed for testing stool was used for diagnosis of C. jejuni in the synovial fluid</p><p>[105]. Determination of in vitro antimicrobial susceptibility remains controversial, since</p><p>most infections are self‐limited, but it should be performed in extraintestinal or severe</p><p>cases. C. jejuni is susceptible to macrolides, fluoroquinolones, aminoglycosides, tetracy-</p><p>clines, and chloramphenicol. C. jejuni is resistant to trimethoprim and most β‐lactam</p><p>antibiotics [106]. Of concern is the increased incidence of resistance to antimicrobials in</p><p>campylobacters, particularly to fluoroquinolones [106–108].</p><p>Therapy</p><p>In general, surgical treatment of Campylobacter PJI is dictated by the clinical presenta-</p><p>tion, underlying conditions, and the status of the prosthetic device. Some cases were</p><p>treated with DAIR, followed by chronic oral suppression. One patient with C. lari bacte-</p><p>remia died of septic complications [100]. Other patients were managed with antimicro-</p><p>bial therapy without surgery. One of them received one month of chloramphenicol and</p><p>died after two months of an unrelated cardiac cause [109]. The second patient received</p><p>eight weeks of parenteral antimicrobials (main treatment erythromycin and ciprofloxa-</p><p>cin), and had a good outcome at six months of follow‐up [102].</p><p>34 Bone and Joint Infections</p><p>Erysipelotrix rhusiopathiae</p><p>E. rhusiopathiae is the causative organism of erysipeloid. It is associated with exposure to</p><p>domestic swine, muskox die‐offs, and linked to the emergence of a new disease syndrome</p><p>in Arctic fox. Several cases of TKA and THA PJI have been reported [110–114]. All but</p><p>one case had exposure to animals, owned a hunting dog, or was a butcher. Two patients</p><p>were immunocompromised due to steroid use, rheumatoid arthritis, lupus nephritis, or</p><p>chronic alcoholism. Systemic symptoms are generally absent.</p><p>Diagnosis relies on isolation of microorganism in culture. Molecular diagnosis includes</p><p>16S rRNA sequencing. PCR on joint fluid may be a sensitive and specific way to detect</p><p>the microorganism, particularly in cases with high index of suspicion but negative cul-</p><p>tures [111].</p><p>E. rhusiopathiae is intrinsically resistant to vancomycin and aminoglycosides. It is sus-</p><p>ceptible to penicillins, broad spectrum cephalosporins, and fluorquinolones [115].</p><p>Treatment requires a combination of surgical intervention and antimicrobial</p><p>therapy.</p><p>The surgical treatment involved resection arthroplasty, TSE in the reported cases.</p><p>Anaerobic Microorganisms</p><p>Clostridium difficile</p><p>C. difficile rarely causes extra‐colonic disease and it is an unusual cause of bone and joint</p><p>infection. Clostridium PJI typically occurs in patients with underlying gastrointestinal</p><p>disease. C. difficile has been identified in hip, knee, and shoulder PJIs [116–118]. Previous</p><p>antimicrobial use was reported in all cases of C. difficile THA. Two cases presented as</p><p>late chronic infections and occurred after C. difficile‐associated diarrhea (CDAD) at vari-</p><p>able intervals. Mc Carthy et al. [119] reported a case of C. difficile THA PJI that occurred</p><p>12 months after resolution of CDAD. The strain isolated from the THA was identical to</p><p>the one isolated from stool 12 months earlier using pulse‐field gel electrophoresis.</p><p>Therapy</p><p>All reported cases were treated with metronidazole. Resection arthroplasty and limb amputa-</p><p>tion after failure of open arthrotomy and antimicrobial therapy were required in two patients,</p><p>respectively. Death occurred in one case as a result of complications of primary disease.</p><p>Actinomyces</p><p>Actinomyces spp. rarely cause PJIs. Comorbid conditions included obesity, IV‐drug use,</p><p>diabetes mellitus, dental procedures, and IUD [120–129]. Patients presented with late</p><p>onset symptoms following primary TJA or revision surgery, the interval ranging from 20</p><p>days and 11 years. Both monomicrobial and polymicrobial infections have been reported.</p><p>Microbiologic Diagnosis</p><p>Actinomyces spp. is difficult to culture, and in the lab, growth can take 5–20 days [130].</p><p>16S rRNA gene sequencing may help with detection of Actinomyces. The Matrix‐Assisted</p><p>Laser Desorption Ionization‐Time of Flight Mass Spectrometry (MALDI‐TOF) can</p><p>provide rapid and accurate identification [124].</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 35</p><p>Therapy</p><p>Among reported cases, most patients were treated with TSE, but resection arthroplasty,</p><p>DAIR [121], or OSE have been reported [128]. Following reimplantation, some authors</p><p>have suggested a six‐week to three‐month course of amoxicillin or oral penicillin [124].</p><p>Actinomyces spp. are susceptible to beta‐lactams, but oxacillin, cloxacillin, and cephalexin</p><p>are not effective. Doxycycline has been used in cases of penicillin allergy. Reported cases of</p><p>Actinomyces PJIs were treated for six to nine weeks with oral or parenteral beta‐lactams,</p><p>followed by four weeks to one year, or even indefinite, oral suppressive therapy [120,121,123].</p><p>Other Anaerobes</p><p>Cases of PJI caused by Veionella dispar, V. parvula, Prevotella melaninogenica, and</p><p>Clostridium perfringens were reported or included in PJI series [131–134]. Clostridium</p><p>septicum PJI has been associated with intestinal malignancy [135,136].</p><p>Mycobacteria</p><p>Mycobacterium tuberculosis</p><p>Infection of the musculoskeletal system represents 1–5% of tuberculosis cases [137–139].</p><p>Few case series of M. tuberculosis total joint arthroplasty infections have been reported.</p><p>In a retrospective study on 2116 episodes of PJI over a 22‐year period, only 7 (0.3%) were</p><p>due to M. tuberculosis [139,140]. Tuberculous PJI usually involves hip, knee, or shoulder,</p><p>and can result most commonly from either local reactivation or occasionally from</p><p>hematogenous spread. M. tuberculosis TJA infection in patients without prior history of</p><p>tuberculosis has been reported [139,141,142]. The risk of reactivation of M. tuberculosis</p><p>in patients undergoing THA or TKA for quiescent tuberculous native septic arthritis var-</p><p>ies between 0–31%. It is higher for patients receiving a TKA (27%) than for those receiv-</p><p>ing a THA (6%) [140–142]. The use of antituberculous therapy at the time of arthroplasty</p><p>for latent M. tuberculosis septic arthritis may be reasonable for patients who have not</p><p>received modern antimycobacterial chemotherapy. Alternatively, preoperative isoniazid</p><p>prophylaxis for their previous M. tuberculosis septic arthritis could be considered [141–145].</p><p>The duration of prophylaxis remains unknown. In patients with an underlying immunode-</p><p>ficiency, history of tuberculous infection, or with tuberculosis risk factors postoperative</p><p>tuberculosis prophylaxis may be advisable [146]. Obtaining mycobacterial cultures at the</p><p>time of arthroplasty should be done in these cases. PPD skin testing should be performed</p><p>prior to total joint arthroplasty in patients from a high prevalence area, in those who have</p><p>a history of native joint septic arthritis due to an unknown pathogen, or when the underly-</p><p>ing joint disease is unknown. The majority of patients with M. tuberculosis PJI are PPD</p><p>positive, but a negative test has been reported by Tokumoto et al. [147]. Quantiferon‐B</p><p>Gold, can be helpful, but it cannot distinguish between latent and active tuberculosis.</p><p>The clinical course of tuberculous PJI can present in two patterns. First, patients are</p><p>recognized at the time of arthroplasty based on histologic or microbiologic evidence of</p><p>M. tuberculosis infection; second, tuberculosis is only recognized in the late postoperative</p><p>period (>6 weeks). In the latter situation, M. tuberculosis PJI often presents insidiously,</p><p>over weeks to months. A draining sinus is commonly seen and was present in all cases</p><p>described by Berbari et al. [140].</p><p>36 Bone and Joint Infections</p><p>Microbiologic Diagnosis</p><p>Currently, the gold standard for diagnosis includes a joint fluid or synovial fluid analysis</p><p>for acid‐fast bacilli culture and histopathology [146]. However, negative findings on histo-</p><p>pathological specimens or cultures do not necessarily prove the absence of tuberculous</p><p>infection [143,146]. PCR testing for extra pulmonary specimens has a sensitivity between</p><p>53.7% and 100%, depending on sample selection [146]. Neogi et al. [139] reported on a</p><p>73‐year‐old female with tuberculous PJI 14 years following a TKA who had negative syn-</p><p>ovial fluid and joint fluid cultures, but had a positive synovial tissue M. tuberculosis PCR.</p><p>Therapy</p><p>In some cases, a coinfecting bacterial pathogen was reported [139,140]. In vitro suscepti-</p><p>bility testing should be performed for all M. tuberculosis isolates, because of the emer-</p><p>gence of resistance. Initial therapy should include isoniazid, rifampin, and pyrazinamide,</p><p>with the addition of ethambutol or streptomycin in case of suspected isoniazid resistance</p><p>[139,148] (see Chapter 19).</p><p>The optimal medical and surgical therapy for M. tuberculosis PJI is unknown.</p><p>Prolonged antituberculous regimen including rifampin for a median duration of 12</p><p>months in the literature, 14 months in a French multicenter study, can be curative, even</p><p>in the absence of surgery, as the capacity of M. tuberculosis for adherence and biofilm</p><p>formation is much lower as compared to that of staphylococci [149]. For patients with</p><p>late onset M. tuberculosis PJI, medical treatment alone is usually unsuccessful, and</p><p>removal of the prosthesis is often required. TSE [140,149–151], partial OSE [140,149–</p><p>153], DAIR [140,141,147,149,154,155], and medical management alone have all been per-</p><p>formed. Failure occurred in 50% of the patients treated with DAIR, indicating that</p><p>exchange of the device should be favored [140,141,147,149,154,155].</p><p>Mycobacterium bovis</p><p>Cases of Mycobacterium bovis prosthetic joint infection have been reported, mainly as</p><p>complication of Bacillus‐Calmette‐Guerin instillation in patients with bladder cancer</p><p>[156-163].</p><p>Non‐tuberculous mycobacteria</p><p>Mycobacterium chelonae and fortuitum have been rarely described as a cause of PJI.</p><p>These rapid growing mycobacteria are non‐pigmented mycobacteria found usually in soil</p><p>or water. Several articles reported M. fortuitum PJI involving both hips and knees [157–</p><p>169], but very few described M. chelonae PJI [165,170,171] and M. abscessus PJI [172–</p><p>179]. Most of the PJIs caused by M. fortuitum occurred in the early postoperative period.</p><p>M. chelonae usually manifests as late infection. The presentation is usually acute, with</p><p>drainage, abscess,</p><p>and fistulae formation.</p><p>Microbiologic Diagnosis</p><p>Mycobacterium chelonae grows more slowly than common bacteria but more rapidly</p><p>than other mycobacteria on agar plates (5–7 days). Often cultures of M. chelonae are</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 37</p><p>reported as negative, or can be misidentified as a coryneform. M. chelonae could also be</p><p>misidentified as Nocardia asteroides. Primary cultures should be incubated for six weeks</p><p>because original growth may not be observed until several weeks of incubation.</p><p>Differentiation of M chelonae from M. fortuitum by biochemical means is important</p><p>because of their different drug susceptibilities. Susceptibility testing by disk method has</p><p>a definite time advantage over the iron uptake test and appears to be equally accurate</p><p>[179,180]. The recommended susceptibility testing method for rapid growing mycobacte-</p><p>ria is the broth microdilution technique, with MIC determinations and resistance break-</p><p>points similar to those used for other bacterial species (NCCLS).</p><p>Therapy</p><p>M. fortuitum is much more drug susceptible than M. chelonae ssp. abscessus. Essentially</p><p>all isolates have in vitro susceptibility to achievable serum levels of amikacin, cefoxitin,</p><p>imipenem, sulfonamides, fluoroquinolones [179,181]. M. fortuitum isolates may develop</p><p>resistance in spite of multidrug regimens, therefore repeating susceptibility testing during</p><p>treatment is important when failure is suspected [179,182].</p><p>M. abscessus is a very difficult microorganism to control, because it is highly resistant to</p><p>antibiotics. It is also resistant to common disinfectants [183]. It is usually susceptible to</p><p>amikacin and cefoxitin [181,183], and some isolates are moderately susceptible to imipenem.</p><p>Unlike M. fortuitum, isolates of M. chelonae ssp. abscessus are usually resistant in vitro to</p><p>most of the potentially useful oral antimicrobial agents (doxycycline, erythromycin, and</p><p>sulfisoxazole) [183–186]. Linezolid and tedizolid seem to be promising agents against</p><p>M. fortuitum, chelonae, and some M. abscessus isolates. Proposed MIC breakpoints for</p><p>linezolid are ≤8 μg/ml for susceptible strains, 16 μg/ml for moderately susceptible, and 32 μg/</p><p>ml, respectively, for resistant strains [183,187,188]. Linezolid‐resistant strains of M. absces-</p><p>sus have been reported [177]. Tigecycline may also be useful for treatment of M abscessus</p><p>PJIs. Clofazimine is a second line alternative for these infections.</p><p>Antimicrobial therapy alone or in combination with DAIR has been proved to be inef-</p><p>fective in PJI caused by M fortuitum. Most patients required removal of the prosthesis or</p><p>arthrodesis. Prolonged parenteral antimicrobial therapy for six weeks, followed by three</p><p>to six months of oral therapy with bacteriologic evidence of complete elimination of the</p><p>infection before reimplantation, has been suggested [167,177]. M. chelonae TKA infec-</p><p>tion required six weeks of cefoxitin and amikacin, followed by administration of tri-</p><p>methoprim‐sulfamethoxazole for a total of three months of therapy. OSE and chronic</p><p>ciprofloxacin suppression had a good outcome after two‐year follow‐up [170,177].</p><p>Removal of the implant is essential in M. abscessus PJI as this microorganism can display</p><p>a smooth forming phenotype that can be a significant impediment in achieving a micro-</p><p>biological cure [189]. Cases of M. abscessus were treated with resection arthroplasty</p><p>[172,173,179,189], DAIR [179], TSE [175,183], or arthrodesis [183]. Antimicrobial regi-</p><p>mens are complex and directed by the susceptibility data. The recommended duration of</p><p>therapy is 6–12 months. A TSE combined with negative aspiration cultures following at</p><p>least two months of antimicrobial holiday was recommended before reimplantation for</p><p>M abscessus PJI [175]. The role of amikacin‐impregnated cement spacer for M. abscessus</p><p>PJI is debatable. Although the incorporation of antibiotic cement has emerged as the</p><p>standard of care in the treatment of PJIs, the efficacy of this adjunct has not been cor-</p><p>roborated by clinical trials. In our experience, the outcomes of M. abscessus PJI have</p><p>been poor, though a successful outcome has been reported [172,175,177,183].</p><p>38 Bone and Joint Infections</p><p>Mycobacterium avium complex</p><p>Mycobacterium avian complex (MAC) is rarely associated with PJIs. The infection tends</p><p>to occur in immunocompromised hosts (HIV, transplantation, SLE, rheumatoid arthri-</p><p>tis) [190–194]. In contrast to M. tuberculosis PJI, Mycobacterium avium complex PJI</p><p>occurs as a result of recent hematogenous dissemination, as opposed to reactivation.</p><p>McLaughlin et al. [192] described a case of M. avium complex THA infection in a 20‐</p><p>year‐old man with AIDS who also had M. avium complex bacteremia and histopatho-</p><p>logical evidence of disseminated M. avium complex infection in both of his THAs.</p><p>Microbiologic Diagnosis</p><p>Mycobacterium avium complex can be a contaminant which complicates the interpreta-</p><p>tion of the culture result [195]. On pathology, granulomas are poorly formed and host</p><p>inflammatory response is minimal [196].</p><p>Therapy</p><p>Because of the rarity of MAC PJI, optimal management is not clearly defined.</p><p>Susceptibilities should be obtained to guide the treatment. A three‐drug regimen of a</p><p>macrolide, ethambutol, and a rifamycin is recommended for 6–12 months, in addition to</p><p>resection arthroplasty with debridement [193,197]. Resection arthroplasty followed by</p><p>appropriate antimycobacterial therapy provides the best outcome. If DAIR is performed,</p><p>chronic oral suppression may needed to prevent relapse [173,191,193].</p><p>Other Microorganisms</p><p>Mycoplasma</p><p>M. hominis, M. pneumonia, and M. salivarium were reported as rare causes of PJI [198–</p><p>202]. M. hominis PJI should be suspected in patients with clinically infected joint with</p><p>purulent aspirate, negative Gram stain, and negative standard cultures. Mycoplasma</p><p>should be considered in the differential diagnosis in patients with culture‐negative PJIs,</p><p>particularly in those with hypogammaglobulinemia. A case of Mycoplasma hominis of a</p><p>prosthetic knee after transurethral biopsy in an immunocompetent patient was recently</p><p>described [202].</p><p>The microbiology laboratory should be asked to look specifically for M. hominis. Gram</p><p>stain of the joint aspirate is unrevealing. Serum and joint fluid antibody levels become</p><p>detectable or rise during the course of the illness. 16S rRNA sequencing identified</p><p>Mycoplasma hominis in a prosthetic knee [202]. Metagenomic shotgun sequencing detected</p><p>Mycoplasma salivarium in a patient with TKA PJI and hypogammmaglobulinemia [199].</p><p>Methods available for susceptibility testing of M. hominis are not standardized, and do not</p><p>correlate with the clinical outcome. Many strains are susceptible in vitro to tetracyclines,</p><p>clindamycin and are moderately susceptible to rifampin. M. hominis is generally resistant to</p><p>aminoglycosides, beta‐lactam antibiotics, vancomycin, sulfonamides, trimethoprim, and</p><p>erythromycin. Fluoroquinolones are usually active in vitro against M. hominis, but resist-</p><p>ance can be induced in vitro by exposing the microorganism to increasing concentrations</p><p>of fluoroquinolones. Limited information suggests that M. hominis is susceptible in vitro to</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 39</p><p>linezolid and quinupristin‐dalfopristin although no clinical trials have been undertaken</p><p>using these agents in M. hominis infections [203].</p><p>Ureplasma</p><p>U. urealyticum, parvum have been recently described as a rare cause of THA and TKA</p><p>PJI and diagnosed by 16S rRNA sequencing [204–206].</p><p>Echinococcus</p><p>Echinococcal THA PJI was reported in three patients [207–209]. Serologic tests may help</p><p>confirm a suspected case of echinococcosis. A negative serologic test does not rule out</p><p>echinococcosis. The Casoni intradermal test and the Weinberg complement fixation test</p><p>confirmed the diagnosis in one reported case [207]. Pathology made the diagnosis in the</p><p>second case [208].</p><p>The management of bone infestation by Echinococcus</p><p>requires complete resection of</p><p>the involved area. Adjunctive chemotherapy before and after surgery appears to reduce</p><p>the risk of recurrence by inactivating protoscolices and lessening the tension of the cysts</p><p>for easier cyst removal [210]. Removal of the prosthesis and prolonged treatment with</p><p>albendazole is required in cases when complete resection of the cysts is not technically</p><p>feasible.</p><p>Tropheryma whipplei</p><p>PJI caused by T. whipplei has been described in a patient with a TKA two years after</p><p>complete cure of Whipple’s disease in one patient, and in a patient with THA with previ-</p><p>ously undiagnosed Whipple’s disease [211,212]. Microbiologic diagnosis is usually made</p><p>by PCR analysis of the joint fluid. A positive PCR result for small bowel tissue in the</p><p>presence of undiagnosed joint infection may support a diagnosis of T. whipplei joint</p><p>infection. Attempts to culture the causative organism are unsuccessful. DAIR of the knee</p><p>prosthesis and chronic oral antimicrobial suppression with trimethoprim‐sulfamethoxa-</p><p>zole followed by pristinamycin controlled the infection after one year of follow‐up [211].</p><p>The patient with THA PJI was treated with OSE followed by nine months of trimetho-</p><p>prim sulfamethoxazole with good short‐term outcome [212].</p><p>Borrelia burgdorferi</p><p>TKA PJI caused by B. burgdorferi has been described in four patients [213–215]. All cases</p><p>occurred in an endemic area. A history of tick exposure was not always apparent. The</p><p>onset of symptoms was acute in all presented cases. The diagnosis for these culture</p><p>negative PJI was made by Lyme serology and Lyme PCR in the synovial fluid. Optimal</p><p>therapy remains unknown, but four to six weeks of antimicrobial therapy (ceftriaxone,</p><p>doxycycline for the reported cases) seemed to be adequate. Surgical intervention entailed</p><p>DAIR in two patients, TSE and no surgery, respectively, in two other patients [213–215].</p><p>In regions with high prevalence of Lyme disease, synovial fluid Lyme PCR testing in</p><p>conjunction with Lyme antibody testing should be done, particularly in cases of culture‐</p><p>negative PJI.</p><p>40 Bone and Joint Infections</p><p>Key Points</p><p>● When treating patients with PJI, an appropriate exposure history is of great clini-</p><p>cal interest, because adequate antimicrobial therapy depends on the correct etiologic</p><p>diagnosis. This is especially important in situations, when routine bacterial cultures fail</p><p>to identify a microorganism despite proof of PJI with non‐microbiological criteria.</p><p>● In patients with culture‐negative PJI, it is crucial to maintain a high index of suspi-</p><p>cion, because unusual or fastidious microorganisms require special stains and culture</p><p>conditions, or modern molecular methods (see Chapter 4).</p><p>● Communication between the microbiologist and the orthopedic infectious disease</p><p>specialist is extremely important for final identification of these microorganisms.</p><p>● Surgical management of PJI‐patients caused by rare microorganisms cannot be stand-</p><p>ardized, because no large series with defined treatment are published.</p><p>References</p><p>1. Askar M, Bloch B, Bayston R. Small‐colony variant of Staphylococcus lugdunensis in</p><p>prosthetic joint infection. Arthroplast Today. 2018;4(3):257–260.</p><p>2. Brandt CM, Duffy MC, Berbari EF, et al. Staphylococcus aureus prosthetic joint infec-</p><p>tion treated with prosthesis removal and delayed reimplantation arthroplasty. Mayo Clin</p><p>Proc. 1999;74(6):553–558.</p><p>3. Sampathkumar P, Osmon DR, Cockerill FR, 3rd. Prosthetic joint infection due to</p><p>Staphylococcus lugdunensis. Mayo Clin Proc. 2000;75(5):511–512.</p><p>4. Weightman NC, Allerton KE, France J. Bone and prosthetic joint infection with</p><p>Staphylococcus lugdunensis. J Infect. 2000;40(1):98–99.</p><p>5. Nagy MT, Hla SM, Keys GW. Late Streptococcus bovis infection of total knee replace-</p><p>ment complicated by infective endocarditis and associated with colonic ulcers. BMJ Case</p><p>Rep. 2013;2013.</p><p>6. Thompson JC, Goldman AH, Tande AJ, et  al. Streptococcus bovis Hip and Knee</p><p>Periprosthetic Joint Infections: A Series of 9 Cases. J Bone Jt Infect. 2020;5(1):1–6.</p><p>7. Wilde AH, Ruth JT. Two‐stage reimplantation in infected total knee arthroplasty. Clin</p><p>Orthop. 1988(236):23–35.</p><p>8. Emerton ME, Crook DW, Cooke PH. Streptococcus bovis‐infected total hip arthro-</p><p>plasty. J Arthroplasty. 1995;10(4):554–555.</p><p>9. Clarridge JE, 3rd, Attorri SM, Zhang Q, et al. 16S ribosomal DNA sequence analysis distin-</p><p>guishes biotypes of Streptococcus bovis: Streptococcus bovis Biotype II/2 is a separate genospecies</p><p>and the predominant clinical isolate in adult males. J Clin Microbiol. 2001;39(4):1549–1552.</p><p>10. Eggelmeijer F, Petit P, Dijkmans BA. Total knee arthroplasty infection due to Gemella haemol-</p><p>ysans. Br J Rheumatol. 1992;31(1):67–69.</p><p>11. von Essen R, Ikavalko M, Forsblom B. Isolation of Gemella morbillorum from joint fluid.</p><p>Lancet. 1993;342(8864):177–178.</p><p>12. Leung DT, Davis EM, Quian Q et al. First report of prosthetic joint infection by Gemella san-</p><p>guinis and associated “pseudosatelliting” phenomenon on culture. J Clin Microbiol.</p><p>2011;49(9):3395–3397.</p><p>13. von Essen R. Blood culture bottles improve microbe isolation rate from joint fluids. Am J Clin</p><p>Pathol. 1989;91(4):501–503.</p><p>14. Nieman RE, Lorber B. Listeriosis in adults: a changing pattern. Report of eight cases and</p><p>review of the literature, 1968‐1978. Rev Infect Dis. 1980;2(2):207–227.</p><p>15. Bush LM, Abdulah A, Perez MT. Listeria monocytogenes prosthetic joint infections. A review</p><p>a propos a case report. 2014; Infect Dis. Clin Pract: 1–4.</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 41</p><p>16. Allerberger F, Kasten MJ, Cockerill FR, 3rd, et  al. Listeria monocytogenes infection in</p><p>prosthetic joints. Int Orthop. 1992;16(3):237–239.</p><p>17. Diaz‐Dilernia F, Costantini J, Nicolino TI et al. Unusual Listeria monocytogenes hematoge-</p><p>nous infection in total knee replacement treated with one‐stage revision surgery. Arthroplasty</p><p>today 2019:296–300.</p><p>18. Seo Y, Noh YS, Wie SH, et al. Prosthetic knee joint infection due to Listeria monocytogenes</p><p>bacteremia in a diabetic female. Korean J Intern Med. 2016 31:616–619.</p><p>19. Tabib W, Guiffault P, Lemort CB, et al. Prosthetic hip joint infection caused by Listeria mono-</p><p>cytogenes. Acta Orthop Belg. 2002;68(2):182–186.</p><p>20. Arnal C, Man H, Delisle F, M’Bappe P, et al. Nocardia infection of a joint prosthesis compli-</p><p>cating systemic lupus erythematosus. Lupus. 2000;9(4):304–306.</p><p>21. Robinson D, Halperin N. Nocardia asteroides infection of an Austin‐Moore hemiarthroplasty</p><p>in a nonimmunocompromised host A case report. Bull Hosp Jt Dis Orthop Inst. 1989;</p><p>49(1):107–110.</p><p>22. Hadeed MM, MacDonell JR, Dempsey IJ et al. Chronic Nocardia cyriacigeorgica Periprosthetic</p><p>Knee Infection Successfully Treated with a Two‐Stage Revision: A Case Report. JBJS Case</p><p>Connect. 2017;7(4):e74.</p><p>23. Laurent F, Rodriguez‐Villalobos H, Cornu O, et al. Nocardia prosthetic knee infection success-</p><p>fully treated by one‐stage exchange: case report and review. Acta Clin Belg. 2015;70(4):</p><p>287–290.</p><p>24. Palmer DL, Harvey RL, Wheeler JK. Diagnostic and therapeutic considerations in Nocardia</p><p>asteroides infection. Medicine. 1974;53(5):391–401.</p><p>25. Ambaye A, Kohner PC, Wollan PC, et al. Comparison of agar dilution, broth microdilution,</p><p>disk diffusion, E‐test, and BACTEC radiometric methods for antimicrobial susceptibility test-</p><p>ing of clinical isolates of the Nocardia asteroides complex. J Clin Microbiol. 1997;35(4):</p><p>847–852.</p><p>26. Wallace RJ, Jr, Septimus EJ, Williams TW, et al. Use of trimethoprim‐sulfamethoxazole for</p><p>treatment of infections due to Nocardia. Rev Infect Dis. 1982;4(2):315–325.</p><p>27. Pidoux O, Argenson JN, Jacomo V, et al. Molecular identification of a Dietzia maris hip pros-</p><p>thesis infection isolate. J Clin Microbiol. 2001;39(7):2634–2636.</p><p>28. Larkin JA, Lit L, Sinnott J, et al. Infection of a knee prosthesis with Tsukamurella species.</p><p>South Med J. 1999;92(8):831–832.</p><p>29. Harrington RD, Lewis CG, Aslanzadeh J, et al. Oerskovia xanthineolytica infection of a pros-</p><p>thetic</p><p>joint: case report and review. J Clin Microbiol. 1996;34(7):1821–1824.</p><p>30. Funke G, von Graevenitz A, Clarridge JE, 3rd et al. Clinical microbiology of coryneform bac-</p><p>teria. Clin Microbiol Rev. 1997;10(1):125–159.</p><p>31. Reboli AC, Bryan CS, Farrar WE. Bacteremia and infection of a hip prosthesis caused by</p><p>Bacillus alvei. J Clin Microbiol. 1989;27(6):1395–1396.</p><p>32. Ha J, Park YJ, Kim YJ et al. Late prosthetic joint infection and bacteremia by Bacillus cereus</p><p>confirmed by 16S rRNA sequencing and hip joint tissue pathology. Ann Clin Microbiol</p><p>2016;19(2):54–57.</p><p>33. Rae S, Webley M, Snaith ML. Salmonella typhimurium arthritis in rheumatoid disease.</p><p>Rheumatol Rehabil. 1977;16(3):150–151.</p><p>34. Widmer AF, Colombo VE, Gachter A, et al. Salmonella infection in total hip replacement:</p><p>tests to predict the outcome of antimicrobial therapy. Scand J Infect Dis. 1990;22(5):</p><p>611–618.</p><p>35. Samra Y, Shaked Y, Maier MK. Nontyphoid salmonellosis in patients with total hip replace-</p><p>ment: report of four cases and review of the literature. Rev Infect Dis. 1986;8(6):978–983.</p><p>36. Arda B, Sipahi OR, Yamazhan T, et al. Salmonella enteritidis related prosthetic joint infection.</p><p>West Indian Med J. 2006;55(6):454–455.</p><p>37. Ekinci M, Bayram S, Akgul T. et al. Periprosthetic joint infection caused by Salmonella‐ Case</p><p>reports of two azathioprine and prednisolone induced immunocompromised patients. Hip</p><p>Pelvis 2017;29(2):139–144.</p><p>42 Bone and Joint Infections</p><p>38. Gupta A, Berbari EF, Osmon DR et al. Prosthetic joint infection due to Salmonella species: a</p><p>case series. BMC Infect Dis. 2014;14:633.</p><p>39. Lo I‐F, Chang. H‐C. Salmonella Septic Arthritis in A Patient with A Hip Implant: A Case</p><p>Report. Intern J Gerontology 2018;12(4):344–347.</p><p>40. Tsukayama DT, Estrada R, Gustilo RB. Infection after total hip arthroplasty. A study of the</p><p>treatment of one hundred and six infections. J Bone Joint Surg Am. 1996;78(4):512–523.</p><p>41. Hashimoto Y, Itho Y, Fujinaga Y, et  al. Development of nested PCR based on the ViaB</p><p>sequence to detect Salmonella typhi. J Clin Microbiol. 1995;33(3):775–777.</p><p>42. Glynn MK, Bopp C, Dewitt W, et al. Emergence of multidrug‐resistant Salmonella enterica</p><p>serotype typhimurium DT104 infections in the United States. N Engl J Med.</p><p>1998;338(19):1333–1338.</p><p>43. Chiu CH, Wu TL, Su LH, et al. The emergence in Taiwan of fluoroquinolone resistance in</p><p>Salmonella enterica serotype choleraesuis. N Engl J Med. 2002;346(6):413–419.</p><p>44. Day LJ, Qayyum QJ, Kauffman CA. Salmonella prosthetic joint septic arthritis. Clin Microbiol</p><p>Infect. 2002;8(7):427–430.</p><p>45. Chen CM, Lu TC, Lo WH. Salmonella infection in total hip replacement‐‐report of successful</p><p>reimplantation and review of the literature. Chung Hua i Hsueh Tsa Chih ‐ Chinese Med J.</p><p>1999;62(7):472–476.</p><p>46. Widmer AF, Colombo VE, Gächter A, et al. Salmonella Infection in Total Hip Replacement:</p><p>Tests to Predict the Outcome of Antimicrobial Therapy. Scand J Infect Dis.</p><p>1990;22(5):611–618.</p><p>47. Fu TS, Ueng SW. Two‐staged revision total hip arthroplasty due to Salmonella infection: case</p><p>report. Chang Gung Medical Journal. 2001;24(3):202–207.</p><p>48. Tattevin P, Cremieux A, Joly‐Guillou M. First case of Salmonella hirschfeldii infection of a</p><p>prosthetic hip. Clin Microbiol Infect. 1998;4:228–230.</p><p>49. Madan S, Abbas D, Jowett RL, et al. Salmonella enteritidis infection in total knee replacement.</p><p>Rheumatology (Oxford). 2001;40(1):112–113.</p><p>50. Rae S, Webley M, Snaith ML. Salmonella typhymurium arthritis in rheumatoid disease.</p><p>Rheumatol Rehabil. 1977; 16:150–151.</p><p>51. Samra Y, Shaked Y, Maier MK. Nontyphoid Salmonellosis in Patients with Total Hip</p><p>Replacement: Report of Four Cases and Review of the Literature. Rev Infect Dis</p><p>1986.;8(6):978–983.</p><p>52. Vikram HR, Buencamino RB, Aronin SI. Primary meningococcal arthritis in a prosthetic knee</p><p>joint. J Infect. 2001;42(4):279–281.</p><p>53. Carral BB, Manoja EA, Cardenas SL. Neisseria meningitidis infecting a Prosthetic Knee Joint:</p><p>A New Case of an Unusual Disease. Case Rep Infect Dis. 2017: 1–3.</p><p>54. McCarthy A, Broderick JM, AP. M. Neiserria meningitidis as a cause of septic arthritis: an</p><p>unusual case of periprosthetic joint infection. Case Rep Infect Dis. 2020:1–3.</p><p>55. Söderquist B. Prosthetic hip joint infection caused by non‐capsulated Haemophilus influenzae.</p><p>Scand J Infect Dis. 2014;46(9):665–668.</p><p>56. Khan S, Reedy S. Haemophilus influenzae infection is a prosthetic knee joint in a patient with</p><p>CLL: a vaccine preventable disease. BMJ Case Rep. 2013:1–3.</p><p>57. Bezwada HP, Nazarian DG, Booth RE, Jr. Haemophilus influenzae infection complicating a</p><p>total knee arthroplasty. Clin Orthop. 2002(402):202–205.</p><p>58. Borenstein DG, Simon GL. Hemophilus influenzae septic arthritis in adults. A report of four</p><p>cases and a review of the literature. Medicine (Baltimore). 1986;65(3):191–201.</p><p>59. Jellicoe PA, Cohen A, Campbell P. Haemophilus parainfluenzae complicating total hip arthro-</p><p>plasty: a rapid failure. J Arthroplasty. 2002;17(1):114–116.</p><p>60. Manian FA. Prosthetic joint infection due to Haemophilus parainfluenzae after dental surgery.</p><p>South Med J. 1991;84(6):807–808.</p><p>61. Ortega‐Andreu M, Rodriguez‐Merchan EC, Aguera‐Gavalda M. Brucellosis as a cause of sep-</p><p>tic loosening of total hip arthroplasty. J Arthroplasty. 2002;17(3):384–387.</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 43</p><p>62. Rodriguez Zapata M, Gamo Herranz A, De La Morena Fernandez J. Comparative study of</p><p>two regimens in the treatment of brucellosis. Chemioterapia. 1987;6(2 Suppl):360–362.</p><p>63. Weil Y, Mattan Y, Liebergall M, Rahav G. Brucella prosthetic joint infection: a report of 3</p><p>cases and a review of the literature. Clin Infect Dis. 2003;36(7):e81–86.</p><p>64. Agarwal S, Kadhi SK, Rooney RJ. Brucellosis complicating bilateral total knee arthroplasty.</p><p>Clin Orthop. 1991(267):179–181.</p><p>65. Malizos KN, Makris CA, Soucacos PN. Total knee arthroplasties infected by Brucella meliten-</p><p>sis: a case report. Am J Orthop. 1997;26(4):283–285.</p><p>66. Orti A, Roig P, Alcala R, Navarro V, et al. Brucellar prosthetic arthritis in a total knee replace-</p><p>ment. Eur J Clin Microbiol Infect Dis. 1997;16(11):843–845.</p><p>67. Yagupsky P, Peled N, Press J, et al. Rapid detection of Brucella melitensis from blood cultures</p><p>by a commercial system. Eur J Clin Microbiol Infect Dis. 1997;16(8):605–607.</p><p>68. Balkhair A, Al Maskari S, Ibrahim S et al. Brucella Periprosthetic Joint Infection Involving</p><p>Bilateral Knees with Negative Synovial Fluid Alpha‐Defensin. Case Rep Infect Dis. 2019(1–3).</p><p>69. Solís García del Pozo JI, Solera J. Systematic review and meta‐analysis of randomized clinical</p><p>trials in the treatment of human brucellosis. Plos One 2012;7(2), e32090.</p><p>70. Lang R, Rubinstein E. Quinolones for the treatment of brucellosis. J Antimicrob Chemother.</p><p>1992;29(4):357–360.</p><p>71. Jones RB, Smith J, Hoffmann A, et al. Secondary infection of a total hip replacement with</p><p>Brucella abortus. Orthopedics. 1983;6:184–186.</p><p>72. Chrdle A, Trnka T, Musil D et al. Francisella tularensis periprosthetic joint infections diag-</p><p>nosed with growth in cultures. J Clin Microbiol. 2019;57:1–7.</p><p>73. Cooper CL, Van Caeseele P, Canvin J, et al. Chronic prosthetic device infection with Francisella</p><p>tularensis. Clin Infect Dis. 1999;29(6):1589–1591.</p><p>74. Rawal H, Patel A, Moran M. Unusual case of prosthetic joint infection caused by Francisella</p><p>tularensis. UBMJ Case Rep. 2017:1–3.</p><p>75. Koneman EN, Allen SD, Janda WM. Other miscellanous fastidious Gram‐negative bacteria.</p><p>Color atlas and textbook of diagnostic microbiology. 5th ed. Philadelphia: Lipincott‐Raven;</p><p>1997. p. 431.</p><p>76. Iglesias L, Garcia‐Arenzana JM, Valiente A, et al. Yersinia enterocolitica O:3 infection of a</p><p>prosthetic knee joint related to recurrent hemarthrosis. Scand J Infect Dis. 2002;34(2):</p><p>132–133.</p><p>77. Oni JA, Kangesu T. Yersinia enterocolitica infection of a prosthetic knee joint. Br J Clin Pract.</p><p>1991;45(3):225.</p><p>78. Jalava‐Karvinen P, Oksi J, al. Rantakokko‐Jalava K et al. Yersinia enterocolitica infection of</p><p>9 Native Joint Arthritis in Adults 139</p><p>Florian B. Imhoff, David E. Bauer, and Ilker Uçkay</p><p>Introduction 139</p><p>Risk Factors 139</p><p>Pathogenesis, Epidemiology, and Microbiology 140</p><p>Diagnosis 141</p><p>Treatment 144</p><p>Outcome 147</p><p>Key Points 148</p><p>Acknowledgments 148</p><p>References 148</p><p>Chapter 10 Septic Arthritis of Axial Joints 151</p><p>Werner Zimmerli</p><p>Septic Arthritis of the Sternoclavicular Joint 151</p><p>Key Points 157</p><p>Septic Arthritis of the Symphysis Pubis 157</p><p>Key Points 161</p><p>Septic Arthritis of the Sacroiliac Joint 161</p><p>Key Points 165</p><p>References 165</p><p>Chapter 11 Periprosthetic Joint Infection: General Aspects 171</p><p>Werner Zimmerli</p><p>Introduction 171</p><p>Definition 172</p><p>Classification 172</p><p>Pathogenesis 172</p><p>Laboratory Investigation 176</p><p>Therapeutic Management 178</p><p>Prophylaxis 179</p><p>Errors in the Management of PJI 180</p><p>Key Points 180</p><p>References 181</p><p>viii Contents</p><p>Chapter 12 Periprosthetic Joint Infection after Total Hip and</p><p>Knee Arthroplasty 187</p><p>Werner Zimmerli, Rihard Trebse, and Martin Clauss</p><p>Introduction 187</p><p>Risk Factors 188</p><p>Microbiology 188</p><p>Clinical Features 189</p><p>Laboratory Investigation 190</p><p>Imaging Procedures 191</p><p>Management 192</p><p>Instructive Cases 202</p><p>References 206</p><p>Chapter 13 Periprosthetic Joint Infection after Shoulder Arthroplasty 213</p><p>Parham Sendi, Andreas Marc Müller, Beat K. Moor, and Matthias A. Zumstein</p><p>Introduction 213</p><p>Risk Factors 213</p><p>Microbiology 215</p><p>Pathogenesis 216</p><p>Clinical Features 217</p><p>Laboratory Investigation 218</p><p>Imaging Procedures 220</p><p>Management 220</p><p>Instructive Cases 224</p><p>References 226</p><p>Chapter 14 Periprosthetic Joint Infection after Elbow Arthroplasty 231</p><p>Yvonne Achermann, Michael C. Glanzmann, and Christoph Spormann</p><p>Introduction 231</p><p>Microbiology 232</p><p>Clinical Features 232</p><p>Diagnostic Procedures 234</p><p>Management 237</p><p>Instructive Cases 242</p><p>References 245</p><p>Chapter 15 Periprosthetic Joint Infection after Ankle Arthroplasty 251</p><p>Parham Sendi, Bernhard Kessler, and Markus Knupp</p><p>Introduction 251</p><p>Risk Factors 252</p><p>Microbiology 254</p><p>Clinical Features 254</p><p>Laboratory Investigation 255</p><p>Imaging Procedures 256</p><p>Management 257</p><p>Instructive Cases 260</p><p>References 262</p><p>Contents ix</p><p>Chapter 16 Osteomyelitis: Classification 265</p><p>Werner Zimmerli</p><p>Classification According to Pathogenesis 265</p><p>Classification According to the Duration of Infection 267</p><p>Classification According to the Localization 268</p><p>Classification According to the Presence of an Implant 269</p><p>Classification According to Anatomy and Comorbidity 269</p><p>References 270</p><p>Chapter 17 Osteomyelitis in Children 273</p><p>Alexander Aarvold, Priya Sukhtankar, and Saul N. Faust</p><p>Introduction 273</p><p>Epidemiology 273</p><p>Pathophysiology 274</p><p>Clinical Presentation, Diagnosis, and Microbiology 275</p><p>Treatment 282</p><p>Complications 284</p><p>Key Points 285</p><p>References 286</p><p>Chapter 18 Acute Osteomyelitis in Adults 289</p><p>Werner Zimmerli</p><p>Introduction 289</p><p>Pathogenesis 289</p><p>Epidemiology 290</p><p>Microbiology 291</p><p>Risk Factors 293</p><p>Clinical Features 295</p><p>Laboratory Investigation 296</p><p>Imaging Procedures 298</p><p>Clinical and Imaging Differential Diagnosis 300</p><p>Treatment 301</p><p>Key Points 304</p><p>References 304</p><p>Chapter 19 Subacute Osteomyelitis: Tuberculous and Brucellar</p><p>Vertebral Osteomyelitis 309</p><p>Juan D. Colmenero and Pilar Morata</p><p>Introduction 309</p><p>Epidemiology 309</p><p>Clinical Features 310</p><p>Laboratory Investigation 312</p><p>Imaging Procedures 315</p><p>Antimicrobial and Surgical Therapy 318</p><p>Key Points 321</p><p>References 321</p><p>x Contents</p><p>Chapter 20 Chronic Osteomyelitis in Adults 325</p><p>Felix W.A. Waibel, Benedikt Jochum, and Ilker Uçkay</p><p>Introduction 325</p><p>Pathogenesis 326</p><p>Diagnosis 326</p><p>Treatment 328</p><p>Key Points 333</p><p>Acknowledgments 333</p><p>References 333</p><p>Chapter 21 Diabetic Foot Osteomyelitis 337</p><p>Eric Senneville and Olivier Robineau</p><p>Introduction 337</p><p>Classification 337</p><p>Microbiology 339</p><p>Risk Factors 341</p><p>Clinical Features 342</p><p>Inflammatory Parameters 342</p><p>Imaging Procedures 343</p><p>Treatment 344</p><p>Prevention 347</p><p>References 348</p><p>Chapter 22 Osteomyelitis of the Jaws 353</p><p>Werner Zimmerli</p><p>Introduction 353</p><p>Classification 353</p><p>Microbiology 355</p><p>Risk Factors 356</p><p>Clinical Features 356</p><p>Laboratory Investigation 359</p><p>Imaging Procedures 360</p><p>Management 361</p><p>Key Points 363</p><p>References 364</p><p>Chapter 23 Fracture‐Related Infection of the Long Bones 367</p><p>Parham Sendi, Mario Morgenstern, Willem‐Jan Metsemakers, and Martin McNally</p><p>Introduction 367</p><p>Classification and Risk Factors 368</p><p>Microbiology 370</p><p>Clinical Features 371</p><p>Laboratory Investigation 372</p><p>Imaging Procedures 373</p><p>Management 373</p><p>Instructive Cases 380</p><p>References 383</p><p>Contents xi</p><p>Chapter 24 Implant‐Associated Vertebral Osteomyelitis 387</p><p>Todd J. Kowalski and Arick P. Sabin</p><p>Introduction 387</p><p>Classification and Risk Factors 388</p><p>Microbiology 389</p><p>Clinical Features 393</p><p>Diagnostic Procedures 393</p><p>Treatment 395</p><p>Instructive Cases 402</p><p>References 404</p><p>Chapter 25 Postoperative Sternal Osteomyelitis 409</p><p>Parham Sendi, Mihai Constantinescu, and Lars Englberger</p><p>Introduction 409</p><p>Risk Factors 410</p><p>Microbiology 414</p><p>Clinical Features 414</p><p>Laboratory Investigation 415</p><p>Imaging Procedures 417</p><p>Management 417</p><p>Instructive Cases 422</p><p>References 425</p><p>Index 429</p><p>xii</p><p>Alexander Aarvold, MD</p><p>Paediatric Orthopaedics</p><p>University Hospital Southampton</p><p>NHS Foundation Trust</p><p>Southampton, UK</p><p>Yvonne Achermann, MD</p><p>Division of Infectious Diseases and</p><p>Hospital Epidemiology</p><p>University Hospital Zürich</p><p>University of Zürich</p><p>Zürich, Switzerland</p><p>David E. Bauer, MD</p><p>Department of Orthopedic Surgery</p><p>Balgrist University Hospital</p><p>University of Zürich</p><p>Zürich, Switzerland</p><p>Jürgen B. Bulitta, PhD</p><p>Department of Pharmacotherapy and</p><p>Translational Research</p><p>College of Pharmacy</p><p>University of Florida</p><p>Orlando, FL, USA</p><p>Lorenzo Calabro, MD</p><p>QEII Jubilee Hospital Brisbane</p><p>Brisbane, Queensland, Australia</p><p>Martin Clauss, MD</p><p>Centre for Musculoskeletal Infections</p><p>University Hospital Basel</p><p>Department of Orthopaedic and</p><p>Trauma Surgery</p><p>University of Basel</p><p>Basel, Switzerland</p><p>Juan D. Colmenero, MD</p><p>Infectious Diseases Service</p><p>University Regional Hospital</p><p>Malaga, Spain</p><p>Caroline Constant, DMV, MSc, MENG,</p><p>DACV‐LA</p><p>AO Research Institute Davos</p><p>AO Foundation</p><p>Davos, Switzerland</p><p>Mihai Constantinescu, MD</p><p>Department of Plastic</p><p>Reconstructive, and Hand Surgery</p><p>University Hospital of Bern</p><p>University of Bern</p><p>Bern, Switzerland</p><p>Stéphane Corvec, PharmD, PhD</p><p>Clinical Microbiology</p><p>Service de Bactériologie‐Hygiène hospitalière</p><p>Institut de Biologie ‐ CHU de Nantes</p><p>Nantes, France</p><p>List of Contributors</p><p>List of Contributors xiii</p><p>Lars Englberger, MD</p><p>Department of Cardiac Surgery</p><p>Hirslanden Clinic Aarau and Bern</p><p>Bern, Switzerland</p><p>Saul N. Faust, MD</p><p>NIHR Southampton Clinical Research</p><p>Facility</p><p>University Hospital Southampton NHS</p><p>Foundation Trust</p><p>Faculty of Medicine and Institute for</p><p>Life Sciences</p><p>University of Southampton</p><p>Southampton, UK</p><p>Mercedes Gonzalez‐Moreno, MSc</p><p>Charité‐Universitätsmedizin</p><p>Corporate member of Freie Universität</p><p>Berlin</p><p>Humboldt‐ Universität zu Berlin, and</p><p>Berlin Institute of Health</p><p>Center for Musculoskeletal Surgery</p><p>Berlin, Germany</p><p>Florian B. Imhoff, MD</p><p>Department of Orthopedic Surgery</p><p>Balgrist University Hospital</p><p>University of Zurich</p><p>Zürich, Switzerland</p><p>Benedikt Jochum, MD</p><p>Department of Orthopedic Surgery</p><p>Balgrist University Hospital</p><p>University of Zurich</p><p>Zürich, Switzerland</p><p>Bernhard Kessler, MD</p><p>Internal Medicine and Infectious</p><p>Diseases</p><p>Hospital Emmental, Burgdorf</p><p>Burgdorf, Switzerland</p><p>Marcus Knupp, MD</p><p>Mein Fusszentrum</p><p>University of Basel</p><p>Basel, Switzerland</p><p>Todd J. Kowalski, MD</p><p>Department of Internal Medicine</p><p>Gundersen Health System</p><p>University of Wisconsin School of</p><p>Medicine and Public Health</p><p>La Crosse, WI, USA</p><p>Cornelia B. Landersdorfer, PhD</p><p>Centre for Medicine Use and Safety</p><p>Monash Institute of Pharmaceutical</p><p>Sciences</p><p>Monash University (Parkville Campus)</p><p>Melbourne, Australia</p><p>Camelia Marculescu, MD</p><p>Division of Infectious Diseases</p><p>Department of Medicine</p><p>Medical University of South Carolina</p><p>Charleston, SC, USA</p><p>Donara Margaryan, MD</p><p>Charité‐Universitätsmedizin</p><p>Corporate Member of Freie Universität</p><p>Berlin</p><p>Humboldt‐ Universität zu Berlin, and</p><p>Berlin Institute of Health</p><p>Center for Musculoskeletal Surgery</p><p>Berlin, Germany</p><p>Martin McNally, MD</p><p>The Bone Infection Unit</p><p>Nuffield Orthopaedic Centre</p><p>Oxford University</p><p>a</p><p>prosthetic knee joint. Case report and review of the literature on dep sited infections caused by</p><p>Yersinia enterocolitica. Advances Infect Dis. 2013 3:95–99.</p><p>79. Honnorat E, Seng P, Savini H et al. Prosthetic joint infection caused by Pasteurella multocida;</p><p>a case series and review of the literature. BMC Infectious Dis. 2016;16(435):1–7.</p><p>80. Lam PW, Page A. Pasteurella multocida non‐native joint infection after dog lick. A case report</p><p>describing a complicated two ‐stage revision and a comprehensive review of the literature. Can</p><p>J Infect Dis Med Microbiol 2015;26(4):212–217.</p><p>81. Gabuzda GM, Barnett PR. Pasteurella infection in a total knee arthroplasty. Orthop Rev.</p><p>1992;21(5):601, 4–5.</p><p>82. Orton DW, Fulcher WH. Pasteurella multocida: bilateral septic knee joint prostheses from a</p><p>distant cat bite. Ann Emerg Med. 1984;13(11):1065–1067.</p><p>83. Mellors JW, Schoen RT. Pasteurella multocida prosthetic joint infection. Ann Emerg Med.</p><p>1985;14(6):617.</p><p>84. Sugarman M, Quismorio FP, Patzakis MJ. Joint infection by Pasteurella multocida. Lancet.</p><p>1975;2(7947):1267.</p><p>85. Chikwe J, Bowditch M, Villar RN, et al. Sleeping with the enemy: Pasteurella multocida infec-</p><p>tion of a hip replacement. J R Soc Med. 2000;93(9):478–479.</p><p>44 Bone and Joint Infections</p><p>86. Guion TL, Sculco TP. Pasteurella multocida infection in total knee arthroplasty. Case report</p><p>and literature review. J Arthroplasty. 1992;7(2):157–160.</p><p>87. Antuna SA, Mendez JG, Castellanos JL, et al. Late infection after total knee arthroplasty</p><p>caused by Pasteurella multocida. Acta Orthop Belg. 1997;63(4):310–312.</p><p>88. Braithwaite BD, Giddins G. Pasteurella multocida infection of a total hip arthroplasty. A case</p><p>report. J Arthroplasty. 1992;7(3):309–310.</p><p>89. Arvan GD, Goldberg V. A case report of total knee arthroplasty infected by Pasteurella mul-</p><p>tocida. Clin Orthop. 1978(132):167–169.</p><p>90. Spagnuolo PJ. Pasteurella multocida infectious arthritis. Am J Med Sci. 1978;275(3):</p><p>359–363.</p><p>91. Maurer KH, Hasselbacher P, Schumacher HR. Letter: Joint infection by Pasteurella multo-</p><p>cida. Lancet. 1975;2(7931):409.</p><p>92. Maradona JA, Asensi V, Carton JA, et al. Prosthetic joint infection by Pasteurella multocida.</p><p>Eur J Clin Microbiol Infect Dis. 1997;16(8):623–625.</p><p>93. Griffin AJ, Barber HM. Joint infection by Pasteurella multocida. Lancet. 1975;1(7920):</p><p>1347–1348.</p><p>94. Chenouard R, Hoppe E, Lemarie C et al. A rare case of prosthetic joint infection associated</p><p>with Coxiella burnetii. International J Infect Dis. 2019;87:166–169.</p><p>95. Tande AJ, Cunningham SA, Raoult D, et al. A case of Q fever prosthetic joint infection and</p><p>description of an assay for detection of Coxiella burnetii. J Clin Microbiol 2014;51(1):66–69.</p><p>96. Million M, Bellevegue L, Labussiere AS ea. Culture‐negative prosthetic joint arthritis related</p><p>to Coxiella burnetii. Am J Med. 2014;786.</p><p>97. Eldin C, Melenotte C, Mediannikov O. et al. From Q Fever to Coxiella Burnetii Infection:</p><p>A Paradigm Change. Clin Microbiol Rev. 2017;30(1):115–190.</p><p>98. Meriglier E, Sungler A, Elsendoom A. et al. Osteoarticular manifestations of Q fever: A case</p><p>series and literature review. Clin Microbiol Infect. 2018;24:8.</p><p>99. Weisenberg S, Perlada D, Peatman T. Q fever prosthetic joint infection. BMJ Case Rep.</p><p>2017:1–3.</p><p>100. Werno AM, Klena JD, Shaw GM, et al. Fatal case of Campylobacter lari prosthetic joint infec-</p><p>tion and bacteremia in an immunocompetent patient. J Clin Microbiol. 2002;40(3):</p><p>1053–1055.</p><p>101. Dumic I, Sengodan M, Franson J et al. Early onset prosthetic joint infection and bacteremia</p><p>due to Campylobacter fetus subspecies fetus. Case Rep Infect Dis. 2017:1–6.</p><p>102. Peterson MC, Farr RW, Castiglia M. Prosthetic hip infection and bacteremia due to</p><p>Campylobacter jejuni in a patient with AIDS. Clin Infect Dis. 1993;16(3):439–440.</p><p>103. Peterson MC, Farr RW, Castiglia M. Prosthetic hip infection and bacteremia due to</p><p>Campylobacter jejuni in a patient with AIDS. Clin Infect Dis. 1993;16(3):439–440.</p><p>104. Zamora‐López MJ, Álvarez‐García P, García‐Campello M. Prosthetic hip joint infection</p><p>caused by Campylobacter fetus: A case report and literature review. Rev Esp Quimioter.</p><p>2018;31(1):53–57.</p><p>105. Vasoo S, Scwab JJ, Cunningham SA. et al. Campylobacter prosthetic joint infection. J Clin</p><p>Microbiol. 2014;52(5):1771–1774.</p><p>106. Lariviere LA, Gaudreau CL, Turgeon FF. Susceptibility of clinical isolates of Campylobacter</p><p>jejuni to twenty‐five antimicrobial agents. J Antimicrob Chemother. 1986;18(6):681–685.</p><p>107. Endtz HP, Ruijs GJ, van Klingeren B, et al. Quinolone resistance in campylobacter isolated</p><p>from man and poultry following the introduction of fluoroquinolones in veterinary medicine.</p><p>J Antimicrob Chemother. 1991;27(2):199–208.</p><p>108. Tenover FC, Baker CM, Fennell CN et al. Antimicrobial resistance in Campylobacter species.</p><p>In: Nachamkin I, Blaser MJ, Tomkins LS, editors. Campylobacter jejuni: current status and</p><p>future trends. Washington, DC: American Society for Microbiology; 1992. p. 66–73.</p><p>109. Bates CJ, Clarke TC, Spencer RC. Prosthetic hip joint infection due to Campylobacter fetus. J</p><p>Clin Microbiol. 1994;32(8):2037.</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 45</p><p>110. Gazeau P, Rezig S, Quaesaet L, et al. Erysipelothrix rhusiopathiae knee prosthesis infection.</p><p>Med Mal Infect. 2018;48(5):372–373.</p><p>111. Groeschel M, Forde T, Turvey S, et al. An unusual case of Erysipelothrix rhusiopathiae pros-</p><p>thetic joint infection from the Canadian Arctic: whole genome sequencing unable to identify a</p><p>zoonotic source. BMC Infect Dis. 2019;19(1):282.</p><p>112. Hocqueloux L, Poisson DM, Sunder S, et al. Septic arthritis caused by Erysipelothrix rhusi-</p><p>opathiae in a prosthetic knee joint. J Clin Microbiol. 2010;48(1):333–335.</p><p>113. Traer EA, Williams MR, Keenan JN. Erysipelothrix rhusiopathiae infection of a total knee</p><p>arthroplasty an occupational hazard. J Arthroplasty. 2008;23(4):609–611.</p><p>114. Troelsen A, Møller JK, Bolvig L, et al. Animal‐associated bacteria, Erysipelotrix rhusiopathiae,</p><p>as the cause of infection in a total hip arthroplasty. J Arthroplasty. 2010;25(3):497.e21–23.</p><p>115. Brooke CJ, Riley TV. Erysipelothrix rhusiopathiae: bacteriology, epidemiology and clinical</p><p>manifestations of an occupational pathogen. J Med Microbiol. 1999;48(9):789–799.</p><p>116. Al-Tawfiq JA, Babiker MM. Mulit-focal Clostridioides (Clostridium) difficile osteomyelitis in</p><p>a patient with sickle cell anemia: case presentation and literature review. Diagn Microbiol</p><p>Infect Dis. 220;96(1):114915. Doi: 10.1016/j.diagmicrobio.2019.114915.</p><p>117. Pron B, Merckx J, Touzet P, et al. Chronic septic arthritis and osteomyelitis in a prosthetic</p><p>knee joint due to Clostridium difficile. Eur J Clin Microbiol Infect Dis. 1995;14(7):599–601.</p><p>118. Ranganath S, Midturi JK. Unusual case of prosthetic shoulder joint infection due to</p><p>Clostridium difficile. Am J Med Sci. 2013;346(5):422–423.</p><p>119. McCarthy J, Stingemore N. Clostridium difficile infection of a prosthetic joint presenting 12</p><p>months after antibiotic‐associated diarrhoea. J Infect. 1999;39(1):94–96.</p><p>120. Brown ML, Drinkwater CJ. Hematogenous infection of total hip arthroplasty with</p><p>Actinomyces following a noninvasive dental procedure. Orthopedics. 2012;35(7):e1086–1089.</p><p>121. Dagher R, Riaz T, Tande AJ, et al. Prosthetic Joint Infection due to Actinomyces species: A</p><p>case series and review of literature. J Bone Jt Infect. 2019;4(4):174–180.</p><p>122. Dubourg G, Delord M, Gouriet F, et al. Actinomyces gerencseriae hip prosthesis infection: a</p><p>case report. J Med Case Rep. 2015;9:223.</p><p>123. Hedke J, Skripitz R, Ellenrieder M, et al. Low‐grade infection after a total knee arthroplasty</p><p>caused by Actinomyces naeslundii. J Med Microbiol. 2012;61(Pt 8):1162–1164.</p><p>124. Redmond SN, Helms R, Pensiero A. A Case of Actinomyces Prosthetic Hip Infection. Cureus.</p><p>2020;12(7):e9148.</p><p>125. Rieber H, Schwarz R, Krämer O, et  al. Actinomyces neuii subsp. neuii associated with</p><p>periprosthetic infection in total hip arthroplasty as causative agent. J Clin</p><p>Microbiol.</p><p>2009;47(12):4183–4184.</p><p>126. Sharma S, Sharma SC. Forgotten intrauterine contraceptive device ‐ A threat to total hip</p><p>prosthesis: A case report with review of the literature. J Clin Orthop Trauma. 2016;7(2):</p><p>130–133.</p><p>127. Wu F, Marriage NA, Ismaeel A, et al. Infection of a total hip arthroplasty with actinomyces</p><p>israelii: Report of a case. N Am J Med Sci. 2011;3(5):247–248.</p><p>128. Wust J, Steiger U, Vuong H, et al. Infection of a hip prosthesis by Actinomyces naeslundii. J</p><p>Clin Microbiol. 2000;38(2):929–930.</p><p>129. Zaman R, Abbas M, Burd E. Late prosthetic hip joint infection with Actinomyces israelii in</p><p>an intravenous drug user: case report and literature review. J Clin Microbiol. 2002;</p><p>40(11):4391–4392.</p><p>130. Valour F, Sénéchal A, Dupieux C, et al. Actinomycosis: etiology, clinical features, diagnosis,</p><p>treatment, and management. Infect Drug Resist. 2014;7:183–197.</p><p>131. Steckelberg J, Osmon D. Prosthetic joint infections. In: Bisno A, Waldwogel FA, editors.</p><p>Infections associated with indwelling medical devices. Washington, DC: ASM press; 2000. p.</p><p>259–290.</p><p>132. Stern SH, Sculco TP. Clostridium perfringens infection in a total knee arthroplasty. A case</p><p>report. J Arthroplasty. 1988;3(Suppl):S37–40.</p><p>46 Bone and Joint Infections</p><p>133. Marchandin H, Jean‐Pierre H, Carriere C, et al. Prosthetic joint infection due to Veillonella</p><p>dispar. Eur J Clin Microbiol Infect Dis. 2001;20(5):340–342.</p><p>134. Maniloff G, Greenwald R, Laskin R, et al. Delayed postbacteremic prosthetic joint infection.</p><p>Clin Orthop. 1987(223):194–197.</p><p>135. Burnell CD, Turgeon TR, Hedden DR, et al. Paraneoplastic Clostridium septicum infection</p><p>of a total knee arthroplasty. J Arthroplasty. 2011;26(4):666.e9–11.</p><p>136. Economedes DM, Santoro J, Deirmengian CA. Clostridium septicum growth from a total</p><p>knee arthroplasty associated with intestinal malignancy: a case report. BMC Infect Dis.</p><p>2012;12:235.</p><p>137. Farer LS, Lowell AM, Meador MP. Extrapulmonary tuberculosis in the United States. Am J</p><p>Epidemiol. 1979;109(2):205–217.</p><p>138. Watts HG, Lifeso RM. Tuberculosis of bones and joints. J Bone Joint Surg Am.</p><p>1996;78(2):288–298.</p><p>139. Neogi DS, Kumar A, Yadav CS, et al. Delayed periprosthetic tuberculosis after total knee</p><p>replacement: is conservative treatment possible? Acta Orthop Belg. 2009;75(1):136–140.</p><p>140. Berbari EF, Hanssen AD, Duffy MC, et al. Prosthetic joint infection due to Mycobacterium</p><p>tuberculosis: a case series and review of the literature. Am J Orthop. 1998;27(3):219–227.</p><p>141. Spinner RJ, Sexton DJ, Goldner RD, et al. Periprosthetic infections due to Mycobacterium tuber-</p><p>culosis in patients with no prior history of tuberculosis. J Arthroplasty. 1996;11(2):217–222.</p><p>142. Amouyel T, Gaeremynck P, Gadisseux GB, et  al. Mycobacterium tuberculosis infection of</p><p>reverse total shoulder arthroplasty: a case report. J Shoulder Elbow Surg. 2019;28:e271–e274.</p><p>143. Kim YH, Han DY, Park BM. Total hip arthroplasty for tuberculous coxarthrosis. J Bone Joint</p><p>Surg Am. 1987;69(5):718–727.</p><p>144. Eskola A, Santavirta S, Konttinen YT, et al. Arthroplasty for old tuberculosis of the knee.</p><p>Journal of Bone and Joint Surgery ‐ British Volume. 1988;70(5):767–769.</p><p>145. Santavirta S, Eskola A, Konttinen YT, et al. Total hip replacement in old tuberculosis A report</p><p>of 14 cases. Acta Orthop Scand. 1988;59(4):391–395.</p><p>146. Bi AS, Li D, MA Y, et al. Mycobacterium tuberculosis as a cause of periprosthetic joint infec-</p><p>tion after total knee arthroplasty: a review of the literature. Cureus. 2019;11(3):e4325, 1–8.</p><p>147. Tokumoto JI, Follansbee SE, Jacobs RA. Prosthetic joint infection due to Mycobacterium</p><p>tuberculosis: report of three cases. Clin Infect Dis. 1995;21(1):134–136.</p><p>148. American Thoracic Society, CDC, IDSA. Treatment of tuberculosis. MMWR. 2003;52(RR</p><p>11):57.</p><p>149. Uhel F, Corvaisier G, Poinsignon Y, et al. Mycobacterium tuberculosis prosthetic joint infec-</p><p>tions: a case series and literature review. J Infect 2019;78:27–34.</p><p>150. Wolfgang GL. Tuberculosis joint infection following total knee arthroplasty. Clin Orthop.</p><p>1985(201):162–166.</p><p>151. Ueng WN, Shih CH, Hseuh S. Pulmonary tuberculosis as a source of infection after total hip</p><p>arthroplasty A report of two cases. Int Orthop. 1995;19(1):55–59.</p><p>152. Krappel FA, Harland U. Failure of osteosynthesis and prosthetic joint infection due to</p><p>Mycobacterium tuberculosis following a subtrochanteric fracture: a case report and review of</p><p>the literature. Arch Orthop Trauma Surg. 2000;120(7‐8):470–472.</p><p>153. Kreder HJ, Davey JR. Total hip arthroplasty complicated by tuberculous infection.</p><p>J Arthroplasty. 1996;11(1):111–114.</p><p>154. McCullough CJ. Tuberculosis as a late complication of total hip replacement. Acta Orthop</p><p>Scand. 1977;48(5):508–510.</p><p>155. Johnson R, Barnes KL, Owen R. Reactivation of tuberculosis after total hip replacement.</p><p>Journal of Bone and Joint Surgery ‐ British Volume. 1979;61‐B(2):148–150.</p><p>156. Leach WJ, Halpin DS. Mycobacterium bovis infection of a total hip arthroplasty: a case</p><p>report. J Bone Joint Surg Br. 1993;75(4):661–662.</p><p>157. Gomez E, Chiang T, Louie T, et al. Prosthetic Joint Infection due to Mycobacterium bovis after</p><p>Intravesical Instillation of Bacillus Calmette‐Guerin (BCG). Int J Microbiol. 2009;2009:527208.</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 47</p><p>158. Langlois ME, Ader F, Dumistrescu O, et al. Mycobacterium bovis prosthetic joint infection.</p><p>Med Mal Infect. 2016;46(8):445–448.</p><p>159. Metayer B, Menu P, Khatchatourian L, et al. Prosthetic joint infection with pseudo‐tumoral</p><p>aspect due to Mycobacterium bovis infection after Bacillus‐Calmette‐Guerin therapy. Ann</p><p>Phys Rehabil Med. 2018;61(1):62–64.</p><p>160. Nguyen MH, Giordani MM, Thompson GR, 3rd. The double‐edged sword ‐ prosthetic joint</p><p>infection following BCG treatment for bladder cancer: a case report. BMC Infect Dis.</p><p>2019;19(1):331.</p><p>161. Patel A, Elzweig J. Mycobacterium bovis prosthetic joint infection following intravesical instil-</p><p>lation of BCG for bladder cancer. BMJ Case Rep. 2019;12(12).</p><p>162. Storandt M, Nagpal A. Prosthetic joint infection: an extremely rare complication of intrave-</p><p>sicular BCG therapy. BMJ Case Rep. 2019;12(12).</p><p>163. Williams A, Arnold B, Gwynne‐Jones DP. Mycobacterium bovis infection of total hip arthro-</p><p>plasty after intravesicular Bacillus Calmette‐Guérin. Arthroplast Today. 2019;5(4):416–420.</p><p>164. Chazerain P, Desplaces N, Mamoudy P, et al. Prosthetic total knee infection with a bacillus</p><p>Calmette Guerin (BCG) strain after BCG therapy for bladder cancer. J Rheumatol.</p><p>1993;20(12):2171–2172.</p><p>165. Badelon O, David H, Meyer L, et al. Mycobacterium fortuitum infection after total hip pros-</p><p>thesis. A report of 3 cases. Rev Chir Orthop Reparatrice Appar Mot. 1979;65(1):39–43.</p><p>166. Booth JE, Jacobson JA, Kurrus TA, et  al. Infection of prosthetic arthroplasty by</p><p>Mycobacterium fortuitum. Two case reports. J Bone Joint Surg Am. 1979;61(2):300–302.</p><p>167. Herold RC, Lotke PA, MacGregor RR. Prosthetic joint infections secondary to rapidly grow-</p><p>ing Mycobacterium fortuitum. Clin Orthop. 1987(216):183–186.</p><p>168. Horadam VW, Smilack JD, Smith EC. Mycobacterium fortuitum infection after total hip</p><p>replacement. South Med J. 1982;75(2):244–246.</p><p>169. Moerman J, Vandepitte J, Corbeel L, et al. Iatrogenic infections caused by the Mycobacterium</p><p>fortuitum‐chelonei complex. Report of two cases and review. Acta Clin Belg. 1985;40(2):</p><p>92–98.</p><p>170. Pring M, Eckhoff DG. Mycobacterium chelonae infection following a total knee arthroplasty.</p><p>J Arthroplasty. 1996;11(1):115–116.</p><p>171. Heathcock R, Dave J, Yates MD. Mycobacterium chelonae hip infection. J Infect.</p><p>1994;28(1):104–105.</p><p>172. Amit P, Rastogi S, Marya S. Prosthetic knee joint infection due to Mycobacterium abscessus.</p><p>Indian J Orthop. 2017;51(3):337–342.</p><p>173. Eid AJ, Berbari EF, Sia IG, et al. Prosthetic joint infection due to rapidly growing mycobacte-</p><p>ria: report of 8 cases and review of the literature. Clin Infect Dis. 2007;45(6):687–694.</p><p>174. Napaumpaiporn C, Katchamart</p><p>W. Clinical manifestations and outcomes of musculoskeletal</p><p>nontuberculous mycobacterial infections. Rheumatol Int. 2019;39(10):1783–1787.</p><p>175. Nengue L, Diaz MAA, Sherman CE, et  al. Mycobacterium abscessus Prosthetic Joint</p><p>Infections of the Knee. J Bone Jt Infect. 2019;4(5):223–226.</p><p>176. Pace V, Antinolfi P, Borroni E, et al. Treating Primary Arthroprosthesis Infection Caused by</p><p>Mycobacterium abscessus subsp. abscessus. Case Rep Infect Dis. 2019;2019:5892913.</p><p>177. Petrosoniak A, Kim P, Desjardins M, et al. Successful treatment of a prosthetic joint infection</p><p>due to Mycobacterium abscessus. Can J Infect Dis Med Microbiol. 2009;20(3):e94–96.</p><p>178. Wang SX, Yang CJ, Chen YC, et al. Septic arthritis caused by Mycobacterium fortuitum and</p><p>Mycobacterium abscessus in a prosthetic knee joint: case report and review of literature.</p><p>Intern Med. 2011;50(19):2227–2232.</p><p>179. Spanyer J, Foster S, Thum‐DiCesare JA et  al Mycobacterium abscessus: A Rare Cause of</p><p>Periprosthetic Knee Joint Infection. MD Edge Surgery. 2018.</p><p>180. Wallace RJ, Jr, Swenson JM, Silcox VA, et al. Disk diffusion testing with polymyxin and ami-</p><p>kacin for differentiation of Mycobacterium fortuitum and Mycobacterium chelonei. J Clin</p><p>Microbiol. 1982;16(6):1003–1006.</p><p>48 Bone and Joint Infections</p><p>181. Swenson JM, Wallace RJJ, Silcox VA, et al. Antimicrobial susceptibility of five subgroups of</p><p>Mycobacterium fortuitum and Mycobacterium chelonae. Antimicrob Agents Chemother.</p><p>1985;28:807.</p><p>182. Martin ML, Dall L. Emergence of multidrug‐resistant Mycobacterium fortuitum during</p><p>treatment. Chest. 1984;85(3):440–441.</p><p>183. Malhotra R, Kiran B, al. Gautam D, et  al. Mycobacterium abscessus periprosthetic joint</p><p>infection following bilateral total knee arthroplasty. ID Cases. 2019;17:e00542.</p><p>184. Wallace RJ, Jr, Wiss K, Bushby MB, et al. In vitro activity of trimethoprim and sulfamethoxa-</p><p>zole against the nontuberculous mycobacteria. Rev Infect Dis. 1982;4(2):326–331.</p><p>185. Wallace RJ, Jr, Jones DB, Wiss K. Sulfonamide activity against Mycobacterium fortuitum and</p><p>Mycobacterium chelonei. Rev Infect Dis. 1981;3(5):898–904.</p><p>186. Silcox VA, Good RC, Floyd MM. Identification of clinically significant Mycobacterium for-</p><p>tuitum complex isolates. J Clin Microbiol. 1981;14(6):686–691.</p><p>187. Brown‐Elliott BA, Wallace RJ, Jr. Clinical and taxonomic status of pathogenic nonpigmented</p><p>or late‐pigmenting rapidly growing mycobacteria. Clin Microbiol Rev. 2002;15(4):716–746.</p><p>188. Wallace RJ, Jr, Brown‐Elliott BA, Ward SC, et al. Activities of linezolid against rapidly grow-</p><p>ing mycobacteria. Antimicrob Agents Chemother. 2001;45(3):764–767.</p><p>189. Howard ST, Rhoades E, Recht J, et al. Spontaneous reversion of Mycobacterium abscessus</p><p>from a smooth to a rough morphotype is associated with reduced expression of glycopeptidol-</p><p>ipid and reacquisition of an invasive phenotype. Microbiology. 2006;152(Pt 6):1581–1590.</p><p>190. Gupta A, Clauss H. Prosthetic joint infection with Mycobacterium avium complex in a solid</p><p>organ transplant recipient. Transpl Infect Dis. 2009;11(6):537–540.</p><p>191. Ingraham NE, Schneider B, Alpern JD. Prosthetic Joint Infection due to Mycobacterium</p><p>avium‐intracellulare in a Patient with Rheumatoid Arthritis: A Case Report and Review of the</p><p>Literature. Case Rep Infect Dis. 2017;2017:8682354.</p><p>192. McLaughlin JR, Tierney M, Harris WH. Mycobacterium avium intracellulare infection of hip</p><p>arthroplasties in an AIDS patient. J Bone Joint Surg Br. 1994;76(3):498–499.</p><p>193. Tan EM, Marcelin JR, Mason E, et al. Mycobacterium avium intracellulare complex causing</p><p>olecranon bursitis and prosthetic joint infection in an immunocompromised host. J Clin</p><p>Tuberc Other Mycobact Dis. 2016;2:1–4.</p><p>194. Isono SS, Woolson ST, Schurman DJ. Total joint arthroplasty for steroid‐induced osteonecro-</p><p>sis in cardiac transplant patients. Clin Orthop. 1987(217):201–208.</p><p>195. Spinner RJ, Sexton DJ, Vail TP. Mycobacterium avium intracellulare infection. J Bone Joint</p><p>Surg Br. 1995;77(1):165.</p><p>196. Horsburgh CR, Jr. Mycobacterium avium complex infection in the acquired immunodefi-</p><p>ciency syndrome. N Engl J Med. 1991;324(19):1332–1338.</p><p>197. Griffith DE, Aksamit T, Brown‐Elliott BA, et al. An official ATS/IDSA statement: diagnosis,</p><p>treatment, and prevention of nontuberculous mycobacterial diseases. Am J Respir Crit Care</p><p>Med. 2007;175(4):367–416.</p><p>198. Han Z, Burnham C‐A, Clohisy J, et al. Mycoplasma pneumoniae Periprosthetic Joint Infection</p><p>Identified by 16S Ribosomal RNA Gene Amplification and Sequencing: A Case Report. The</p><p>Journal of bone and joint surgery American volume. 2011;93:e103.</p><p>199. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et al. A Novel Prosthetic Joint Infection</p><p>Pathogen, Mycoplasma salivarium, Identified by Metagenomic Shotgun Sequencing. Clin</p><p>Infect Dis. 2017;65(2):332–335.</p><p>200. Sneller M, Wellborne F, Barile MF, et al. Prosthetic joint infection with Mycoplasma hominis.</p><p>J Infect Dis. 1986;153(1):174–175.</p><p>201. Madoff S, Hooper DC. Nongenitourinary infections caused by Mycoplasma hominis in</p><p>adults. Rev Infect Dis. 1988;10(3):602–613.</p><p>202. Rieber H, Frontzek A, Fischer M. Periprosthetic joint infection associated with Mycoplasma</p><p>hominis after transurethral instrumentation in an immunocompetent patient. Unusual or</p><p>underestimated? A case report and review of the literature. Int J Infect Dis. 2019;82:86–88.</p><p>3 Unusual Microorganisms in Periprosthetic Joint Infection 49</p><p>203. Kenny GE, Cartwright FD. Susceptibilities of Mycoplasma hominis, M. pneumoniae, and</p><p>Ureaplasma urealyticum to GAR‐936, dalfopristin, dirithromycin, evernimicin, gatifloxacin,</p><p>linezolid, moxifloxacin, quinupristin‐dalfopristin, and telithromycin compared to their sus-</p><p>ceptibilities to reference macrolides, tetracyclines, and quinolones. Antimicrob Agents</p><p>Chemother. 2001;45(9):2604–2608.</p><p>204. Roerdink RL, Douw CM, Leenders AC, et  al. Bilateral periprosthetic joint infection with</p><p>Ureaplasma urealyticum in an immunocompromised patient. Infection. 2016;44(6):807–810.</p><p>205. Rouard C, Pereyre S, Abgrall S, et al. Early prosthetic joint infection due to Ureaplasma urea-</p><p>lyticum: Benefit of 16S rRNA gene sequence analysis for diagnosis. J Microbiol Immunol</p><p>Infect. 2019;52(1):167–169.</p><p>206. Sköldenberg OG, Rysinska AD, Neander G, et al. Ureaplasma urealyticum infection in total</p><p>hip arthroplasty leading to revision. J Arthroplasty. 2010;25(7):1170.e11–13.</p><p>207. Voutsinas S, Sayakos J, Smyrnis P. Echinococcus infestation complicating total hip replace-</p><p>ment. A case report. J Bone Joint Surg Am. 1987;69(9):1456–1458.</p><p>208. Notarnicola A, Panella A, Moretti L, et al. Hip joint hydatidosis after prosthesis replacement.</p><p>Int J Infect Dis. 2010;14 Suppl 3:e287–290.</p><p>209. Perlick L, Sommer T, Zhou H, et al. Atypical prosthetic loosening in the hip joint. Radiologe.</p><p>2000;40(6):577–579.</p><p>210. Aktan AO, Yalin R. Preoperative albendazole treatment for liver hydatid disease decreases the</p><p>viability of the cyst. Europ J Gastroenterol Hepatol. 1996;8(9):877–879.</p><p>211. Fresard A, Guglielminotti C, Berthelot P, et al. Prosthetic joint infection caused by Tropheryma</p><p>whippelii (Whipple’s bacillus). Clin Infect Dis. 1996;22(3):575–576.</p><p>212. Cremniter J, Bauer T, Lortat‐Jacob A, et al. Prosthetic hip infection caused by Tropheryma</p><p>whipplei. J Clin Microbiol. 2008;46(4):1556–1557.</p><p>213. Adrados M, Wiznia DH, Golden M, et al. Lyme periprosthetic joint infection in total knee</p><p>arthroplasty. Arthroplasty Today. 2018;4(2):158–161.</p><p>214. Collins KA, Gotoff JR, Ghanem ES. Lyme Disease: A Potential Source for Culture‐negative</p><p>Prosthetic Joint Infection. J Am Acad Orthop Surg Glob Res Rev. 2017;1(5):e023.</p><p>215. Wright WF, Oliverio JA. First Case of Lyme Arthritis Involving a Prosthetic Knee Joint. Open</p><p>Forum Infect Dis. 2016;3(2):ofw096.</p><p>51</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>Chapter 4</p><p>Introduction</p><p>Despite the correct implementation</p><p>of special diagnostic culture techniques, such as tis-</p><p>sue sample processing with beadmill, prolonged incubation time, or sonication of</p><p>removed implants, a considerable number of bone and joint infections (BJI) are either</p><p>culture‐negative or misjudged as aseptic failure [1–3]. Misinterpretation may lead to</p><p>wrong or needless antimicrobial treatment, or even to unnecessary surgery [4,5].</p><p>The diagnostic value of non‐culture techniques has been investigated in synovial fluid,</p><p>periprosthetic tissue samples (homogenized manually or by semi‐automated techniques</p><p>[6]), and sonication fluid [7–9]. Although these innovative techniques have proved to be</p><p>helpful, strict conditions have to be respected by specifically trained staff in order to</p><p>avoid any contamination. Different practicable improvements, such as real‐time technol-</p><p>ogy, automated systems, and commercial kits are now available, and support the micro-</p><p>biologists in daily practice [10].</p><p>In this chapter, we describe non‐culture techniques based on nucleic acid amplifica-</p><p>tion, sequencing, and mass‐spectrometry methods. Polymerase chain reaction (PCR) is a</p><p>frequently used technique in most microbiology laboratories that allows detection of a</p><p>nucleic acid fragment by amplifying a sequence. There are different PCR types that can</p><p>be used in the BJI diagnosis, and each one requires the selection of appropriate primers</p><p>(Table 4.1).</p><p>Broad‐Range PCR</p><p>The value of broad‐range PCR has been extensively studied in the diagnosis of BJI, but</p><p>extraction of pathogen DNA from bone, joint tissue, or implant remains challenging [11].</p><p>Identification of Pathogens in Bone</p><p>and Joint Infections by Non‐Culture</p><p>Techniques</p><p>Maria Eugenia Portillo and Stéphane Corvec</p><p>52 Bone and Joint Infections</p><p>Disruption of the biofilm is an essential step to release the DNA in order to improve the</p><p>sensitivity of broad‐range bacterial PCR [12].</p><p>Pathogens can be identified even if genus or species are unknown using universal primers</p><p>to amplify bacterial or fungal DNA, followed by the identification of the species by</p><p>sequencing, a technique which is also called universal PCR. This methodology is applied in</p><p>isolated strains that are difficult to identify using conventional techniques. More recently, it</p><p>has also been used directly from clinical samples, where the detection and identification of</p><p>the pathogen by conventional techniques remains difficult or not possible [11].</p><p>Broad‐range PCR consists of two steps: the amplification of the bacterial or fungal</p><p>DNA within the sample, and the subsequent sequencing of the PCR fragment for the iden-</p><p>tification of the microorganism (Figure 4.1). The regions of the genome that are used must</p><p>fulfill fundamental characteristics. First, they must be present in all bacterial or fungal spe-</p><p>cies; second, they should contain highly conserved sequences to which the primers are</p><p>directed; and finally, they have to include polymorphic sequences, in order to distinguish</p><p>different species. After amplifying and sequencing the fragment, the obtained sequence is</p><p>compared to those placed in public databases such as NCBI GenBank. For sequence align-</p><p>ment, programs such as BLAST are available, and allow online sequence comparison [13].</p><p>In bacteria, species identification at the molecular level is based on analysis of the 16S</p><p>rRNA gene sequence. This molecule is 1,500 bp in size. Usually, it is sufficient to analyze</p><p>the first 500 bp of this gene, since it is the most variable region, but in other cases, the</p><p>entire gene must be sequenced, or even the study of other sites in the genome must be</p><p>used. Similarly, for fungal broad‐range PCR, primers targeting conserved areas of genes</p><p>encoding the 18S, 5.8S, and 28S ribosomal subunits are used. ITS‐1 and 2 regions are</p><p>among these genes, which allow discriminating between the different fungal species [14].</p><p>Therefore, broad‐range PCR allows the identification of microorganisms previously</p><p>not thought to cause infection, despite it being less sensitive than targeted or multiplex</p><p>PCR. The main disadvantages of broad‐range PCR are lack of sensitivity, false‐positive</p><p>results resulting from contamination, need of subsequent sequencing, and challenge of</p><p>result interpretation [15]. Also, the sequence analysis may be uninformative and thus, this</p><p>approach may not indicate whether a polymicrobial or monomicrobial infection is pre-</p><p>sent (i.e., due to overlapping electropherogram peaks) or misleading (i.e., due to missed</p><p>detection of minority species). This issue may be mitigated by more detailed sequence</p><p>Table 4.1. Different types of PCR applied in the diagnosis of BJI.</p><p>PCR General description</p><p>Simultaneous</p><p>amplification and</p><p>detection Quantification</p><p>Broad‐</p><p>range PCR</p><p>Detection of genes universally present</p><p>in microorganisms which requires</p><p>posterior sequencing step for</p><p>identification</p><p>No Semiquantitative by</p><p>comparing gel band</p><p>intensity</p><p>Targeted</p><p>PCR</p><p>Specific detection of a particular</p><p>microorganism and/or resistance</p><p>mechanism</p><p>No: Gel‐based PCR</p><p>Yes: Real‐time PCR</p><p>Semiquantitative</p><p>Yes</p><p>Multiplex</p><p>PCR</p><p>Simultaneously detection of several</p><p>microorganisms and/or resistance</p><p>mechanisms by adding the sets of</p><p>primers of interest</p><p>No: Gel‐based PCR</p><p>Yes: Real‐time PCR</p><p>Semiquantitative</p><p>Yes</p><p>4 Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques 53</p><p>analysis, or through software manipulation of sequencer data (i.e. use of the software</p><p>RipSeq) [16]. During a prospective multicenter cross sectional study, Bémer et al. [17]</p><p>demonstrated a sensitivity of only 73.3% in the diagnosis of prosthetic joint infections</p><p>(PJI). Although new PCR assays have been developed, the sensitivity of PCR for the</p><p>diagnosis of PJI is variable in different studies, whereas its specificity is reliably high,</p><p>allowing the exclusion of PJI [18].</p><p>Targeted PCR</p><p>Targeted or specific PCR can be developed for any known microorganism, and can be</p><p>designed to be extremely sensitive. The analysis is typically performed in real‐time, because</p><p>the amplification process and detection occur simultaneously in the same closed vial.</p><p>Furthermore, it is possible to measure the amount of DNA synthesized at each moment</p><p>during amplification, since the fluorescence emission produced in the reaction is propor-</p><p>tional to the amount of formed DNA. Therefore, real‐time PCR has three important advan-</p><p>tages compared to conventional or agarose gel‐based PCR, namely higher speed, lower</p><p>contamination rate, and more precise DNA quantification. Likewise, point mutations that</p><p>may be related to antimicrobial resistance or virulence factors can also be detected. However,</p><p>if the gene needs to be sequenced, it is better to perform an agarose gel‐based PCR [19,20].</p><p>1500 bp</p><p>1</p><p>Sample Marker</p><p>2</p><p>Figure 4.1. Broad‐range PCR scheme for bacteria and fungi. From the corresponding genome</p><p>region: 1) a DNA fragment is amplified using primers targeting conserved regions present in all</p><p>(universal) bacteria or fungi and 2) polymorphic regions are sequenced 3) to identify the species.</p><p>Source: ABI chromatogram and National Institutes of Health.</p><p>54 Bone and Joint Infections</p><p>Different genes have been used as target to accurately identify most coagulase‐negative</p><p>staphylococci (CNS) implicated in BJI based on PCR amplification and sequencing.</p><p>A genetic classification and identification of Staphylococcus species was conducted based</p><p>on the comparison of partial glyceraldehyde‐3‐phosphate dehydrogenase gene, 16S</p><p>rRNA, hsp60, rpoB, sodA, and tuf gene sequences [21].</p><p>Kingella kingae is currently recognized as the prime etiology of BJI in children aged</p><p>6–48 months (see Chapter 8). The organism is fastidious, its growth is inhibited by syno-</p><p>vial fluid and bone exudates, and its presence in clinical specimens is commonly missed</p><p>by traditional culture methods. Therefore, the use of specific PCR is highly recommended</p><p>in septic arthritis and osteomyelitis in infants and young children [22] (Figure 4.2).</p><p>Different companies have</p><p>developed an automated, easy‐to‐use, fast, and accurate</p><p>real‐time Staphylococcus aureus identification PCR which can be combined with the</p><p>search for methicillin resistance (mecA and/or mecC gene) to optimize the management</p><p>and the appropriate treatment of the patient, especially in acute septic arthritis cases [23].</p><p>The addition of specific nucleic acid amplification testing is also suggested if bacterial</p><p>cultures do not reveal growth or if fastidious microorganisms are suspected. This may typi-</p><p>cally occur in the case of vertebral osteomyelitis (Brucella spp., Mycobacterium tuberculosis,</p><p>Tropheryma whipplei, Coxiella spp., or Mycoplasma spp.) [24] (see Chapters 18 and 19).</p><p>Multiplex PCR</p><p>Multiplex PCR is a technique in which more than two sets of primers are involved in the</p><p>process of amplifying various target sequences, allowing the simultaneous detection and</p><p>identification of different genes. The main advantage of these systems is the ability of</p><p>175 bp</p><p>Molecular</p><p>marker</p><p>A</p><p>Sample</p><p>B</p><p>Figure 4.2. Targeted PCR of cpn60 gene from Kingella kingae detected A) in conventional</p><p>electrophoresis gel and B) in real‐time PCR.</p><p>4 Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques 55</p><p>grouping in a single process different targeted PCR, simplifying the process, saving time</p><p>and cost, as well as shortening the diagnostic time [5].</p><p>In a recent meta‐analysis, it has been shown that multiplex PCR from sonication fluid</p><p>of prosthetic implants is reliable and of great value for the diagnosis of BJI [25]. Several</p><p>groups have investigated different multiplex PCR panels. However, these tests have mainly</p><p>been developed for bloodstream infections. Thus, their use for the rapid diagnosis of BJI</p><p>is off‐label [26–28]. There are different primer panels (Table 4.2), including primers for</p><p>low‐virulent organisms frequently involved in chronic or delayed PJI, such as</p><p>Corynebacterium spp., Cutibacterium spp., or other anaerobes.</p><p>Next‐Generation Sequencing Approach</p><p>Beside culturomics methods, in the last two decades different non‐culture techniques</p><p>have been developed to better detect microorganisms in clinical perioperative samples,</p><p>either bacteria or fungi.</p><p>In order to improve the BJI diagnosis, different teams have developed original sequenc-</p><p>ing approaches. Henceforth, with the reduced cost of sequencing and the availability of</p><p>next‐generation sequencing (NGS) methods, it is reasonable to capture all the potential</p><p>sequences which represent the whole diversity of pathogens present in a joint, a bone</p><p>biopsy, or a tissue in contact with the device (Table 4.3). Interestingly, compared with the</p><p>targeted 16S universal PCR called the metataxonomic method, the shotgun random</p><p>approach allowed the catching of all the sequences including those from fungi or poten-</p><p>tially viruses (Table 4.4). In 2018, Dekker [29] wrote that metagenomics approaches are</p><p>close to reality for the diagnostic workup of clinical infectious diseases. However, we need</p><p>to highlight that the first challenge is a potential microbial DNA contamination, which</p><p>may occur either during sampling, linked to the reagents, due to contaminated instru-</p><p>ment surfaces, or in the environment [30]. Therefore, the interpretation of results needs</p><p>strict criteria and clinical knowledge in order to detect fastidious microorganisms and to</p><p>rule out contaminants. The specialized microbiologists should assist clinicians in inter-</p><p>preting the results of NGS in interdisciplinary meetings. For meaningful use, this new</p><p>tool should only be performed in patients with chronic infection, with culture‐negative</p><p>Table 4.2. Multiplex PCR platforms used in BJI.</p><p>Platform Manufacturer BJI cartridge Target Resistance genes</p><p>Unyvero</p><p>i60</p><p>Curetis Yes GPBa including Corynebacterium spp.,</p><p>Granulicatella spp., and Abiotrophia</p><p>spp., GNBb including non‐fermenting,</p><p>anaerobes, and fungi</p><p>Yes</p><p>SeptiFastc Roche No GPBa, GNBb including non‐fermenting</p><p>and fungi</p><p>Yes</p><p>FilmArray BioFire (BJI) Yes GPBa, GNBb including non‐fermenting,</p><p>Kingella kingae, anaerobes and fungi</p><p>Yes</p><p>aGPB Gram‐positive bacteria</p><p>bGNB Gram‐negative bacteria</p><p>cSoon discontinued</p><p>56 Bone and Joint Infections</p><p>specimens, or with current or recent antimicrobial therapy. Another issue is the frequent</p><p>overrepresentation of human host DNA (neutrophils, osteoblast and osteoclast cells)</p><p>which interferes with the microorganisms’ sequences, less frequent in terms of amount or</p><p>proportion. For these reasons, we should still be reserved with this novel technique.</p><p>However, NGS‐based technology will gradually improve the etiologic diagnosis of BJI.</p><p>They will reach in the near futures the standard of care in infectious disease diagnostics.</p><p>In 2017, Street et al. [31] from the Oxford group published one of the first studies about</p><p>NGS (metagenomic sequencing) on 97 sonication fluid samples from various orthopedic</p><p>devices. They concluded that this technology is a promising method, especially if in the</p><p>future there will be improved techniques to avoid microbial, as well as human DNA,</p><p>contamination. The increasing availability of different systems, which are quicker than</p><p>culture and even portable, will offer a new clinically useful rapid diagnostic tool (Table 4.3).</p><p>However, they also highlighted the risk of false‐positive results, especially with</p><p>Cutibacterium acnes, a bacterium from the skin microbiome, or with environmental bac-</p><p>teria linked to water. In addition, false‐negative results may occur, due to a limited data-</p><p>base. Therefore, each case should be discussed in a multidisciplinary team [31]. During</p><p>the same year, Thoendel et al. [32] from Robin Patel’s group reported the usefulness of</p><p>this approach to identify a wide range of BJI pathogens, including difficult‐to‐detect</p><p>pathogens, especially in culture‐negative infections. Among others, Mycoplasma species,</p><p>anaerobes including Cutibacterium acnes, fungi, Staphylococcus spp., and Streptococcus</p><p>spp. were identified as new microorganisms in culture‐negative BJI. Mixed and polymi-</p><p>crobial PJI could also be detected. However, the data interpretation needs attention.</p><p>Read counts and depth of genome coverage are important to distinguish non‐cultured</p><p>potential microorganisms detected from uninfected cases or negative controls. This con-</p><p>stitutes a major challenge and the background reads should be analyzed carefully to</p><p>determine their relevance.</p><p>Table 4.3. Different applications of Next‐Generation Sequencing (NGS) in clinical microbiology.</p><p>Methodology Applications</p><p>Whole Genome</p><p>Sequencing</p><p>WGS of a pure organism isolated from culture.</p><p>Targeted NGS directly from</p><p>specimen</p><p>16S rDNA from a clinical specimen for bacterial profiling or PCR</p><p>amplification of other specified targets followed by sequencing.</p><p>Metagenomic NGS directly</p><p>from specimen</p><p>Nucleic acid composition of the specimens includes host, microbiome,</p><p>and the maybe the accidental introduction of contaminating nucleic acid.</p><p>Table 4.4. Terms used in the area of Next‐Generation Sequencing.</p><p>Technology Definitions</p><p>Next‐Generation</p><p>Sequencing</p><p>High‐throughput, massively parallel sequencing of DNA fragments</p><p>independently and simultaneously.</p><p>Whole Genome</p><p>Sequencing (WGS)</p><p>Sequencing of a microbial genome. WGS can be applied to pure culture</p><p>growth or directly from specimen.</p><p>Targeted WGS Uses a selection process to enrich for specific targets before sequencing.</p><p>Metagenomic WGS Sequencing of all nucleic acids detected directly from patient specimens.</p><p>4 Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques 57</p><p>During the last couple of years, different case reports or outbreaks have been reported</p><p>in the literature using this new molecular strategy in the orthopedic field [33–37].</p><p>Furthermore, other samples such as synovial fluids have been tested using NGS. The</p><p>metagenomic shotgun sequencing approach can detect microorganisms causing BJI, and</p><p>might be useful especially for negative‐culture cases when applied to synovial</p><p>fluids. Ivy</p><p>et al. [38] described the workflow methodology and interpretation performed during their</p><p>study. They present a list of the most frequent contaminant microorganisms depending</p><p>on the read counts. The read counts went from 2,669 in Cutibacterium to 2,017,563 in</p><p>Acinetobacter. Despite the lack of a generally recognized gold standard in all studies</p><p>analyzing with NGS technology in BJI, it seems important to define a threshold depend-</p><p>ing on the reads detected. Moreover, to limit this bias between pathogens and contami-</p><p>nating microbial DNA, an effective microbial enrichment and DNA isolation remain</p><p>crucial to allow better analysis of the samples. Thus, using a strict protocol and prudent</p><p>interpretation allows us to correctly recognize contaminants.</p><p>In this field, another technology approach is available. Sanderson et al. [39] demon-</p><p>strated the proof of concept using the long‐read sequencing MinION technology on 9</p><p>specimens (7 positive and 2 negative). Despite almost 90% of human DNA, they man-</p><p>aged to detect bacterial infections from DNA extracted directly from sonication fluid</p><p>samples, and potentially provide an answer within minutes of starting sequencing. One</p><p>case was polymicrobial with three different microorganisms, of which Fusobacterium</p><p>nucleatum was not detected by tissue biopsy or sonication fluid culture [39]. Thus, this</p><p>strategy, which is not available in all laboratories, may provide improvement in the diag-</p><p>nostic process in special cases. However, a definition of patient eligibility criteria is</p><p>needed. If used in the correct population, the patient outcome may improve, due to the</p><p>shorter diagnostic delay, which results in a more rapid targeted antibiotic treatment.</p><p>Regarding the analysis of polymicrobial infections using a 16S metagenomic approach,</p><p>Chen et  al. [40] demonstrated that the three databases used provided similar results.</p><p>Indeed, the only variability observed was noticed on less abundant genera, especially</p><p>those for which the relative abundance was less than 5% of the total reads. One limit of</p><p>this recent study is the low number of patients (n=11) and maybe the primers selected, as</p><p>previously mentioned [40].</p><p>At last, with the democratization of NGS (metagenomic and/or shotgun approaches),</p><p>and the decreased costs, it would be interesting to standardize the method used to improve</p><p>the BJI diagnosis, especially using NGS (Figure 4.3). Indeed, as the experimental design</p><p>and analysis methods of each study are different, it would be logical to establish an NGS</p><p>gold standard for PJI constituing a suitable standard procedure for perioperative speci-</p><p>mens, which allows comparing of different studies. Thus, recently Li et al. [41] published</p><p>a review and meta‐analysis of the performance of sequencing assays in the diagnosis of</p><p>PJI. Even if it sometimes remains difficult to compare different studies, sequencing assays</p><p>have the potential to improve biological diagnosis of PJI, especially when culture results</p><p>are negative. However, one should be cautious about which type of culture is performed.</p><p>According to our experience, optimized culture after beadmill processing remains better</p><p>than other methods. The authors highlighted the importance of an antibiotic‐free inter-</p><p>val to improve the performances and enhance the ability to detect microorganisms, espe-</p><p>cially those difficult to culture but susceptible to antibiotics. They also highlighted the</p><p>superiority of sequencing by synthesis. Although Sanger sequencing remains more easily</p><p>available and accurate in sequencing, NGS might be superior in consideration of cost</p><p>and promptness [41]. At last in their model, Torchia et al. [42] suggested that NGS should</p><p>be reserved for patients with a high pretest probability of PJI and specific clinical contexts,</p><p>58 Bone and Joint Infections</p><p>such as antibiotic pretreatment, repetitive previous interventions, risk of polymicrobial</p><p>infection, and chronic infection. We clearly need to define the indications and subgroups</p><p>of patients for which NGS offers clinical benefit [42]. Recently, Wang et al. [43] concluded</p><p>that targeted antibiotic treatment based on NGS results is reliable to identify pathogens</p><p>in patients with culture‐negative PJI, and that it yields a favorable outcome in a short</p><p>period of time. Therefore, we can ask two questions. First, should this technology be</p><p>included in the diagnostic routine procedure? Second, does NGS indeed improve the</p><p>outcome of patients with culture‐negative PJI due to the earlier switch from empiric to</p><p>adapted antibiotic treatment? This may be especially important in patients with fastidi-</p><p>ous microorganisms or those with polymicrobial PJI, because adapted therapy may be</p><p>delayed up to one week in this population, even with optimal culture methods [44].</p><p>Taken together, the NGS approach will be an additional diagnostic test when standard</p><p>of care testing is uninformative. However, in the near future, it will still be limited to spe-</p><p>cialized laboratories, where a sequencing platform (Table 4.5) and a bioinformatic pipe-</p><p>line are available, and the lab workers have the required ability. It is probable that in the</p><p>near future, the costs for this technology will decrease, allowing a broader use of the NGS</p><p>approach.</p><p>Mass Spectrometry‐Based Methods</p><p>The increase of this new technology to identify microorganisms (MALDI‐TOF spec-</p><p>trometry: Matrix Assisted Laser Desorption Ionization – Time of Flight) more than a</p><p>decade ago in microbiology has constituted a real evolution and improvement in the</p><p>turnaround time of identification [45]. With this method, the identification process took</p><p>24h less time, when the culture was positive. Quickly, its use has been diverted to obtain</p><p>a rapid identification directly from samples, in particular blood culture bottles but also</p><p>Figure 4.3. Next‐Generation Sequence approaches, including specimen preparation and</p><p>bioinformatics analysis.</p><p>4 Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques 59</p><p>later in the synovial fluids [46,47]. For more than a decade, the majority of the microbiol-</p><p>ogy labs have been using the MALDI‐TOF technology. Directly from the agar plate, this</p><p>identification method allows a rapid diagnostic (Figure 4.4) [5].</p><p>In 2010, Harris et al. [48] reported the ability and the great potential of the MALDI</p><p>Biotyper software to recognize clonally related strains within a species group (i.e. sub‐</p><p>typing). This MALDI‐TOF/MS MALDI Biotyper system provides a promising rapid and</p><p>reliable method of identifying clinical isolates from BJI to the species level, and has poten-</p><p>tial for sub‐typing [48]. Peel et al. [49] confirmed these data for the diagnosis of PJI. The</p><p>likelihood that a microorganism was a pathogen or contaminant differed with the pros-</p><p>thetic joint location, particularly in the case of Cutibacterium acnes, which is especially</p><p>frequent in shoulder PJI. This spectral method constitutes a valuable tool for the identifica-</p><p>tion of bacteria isolated from prosthetic joints, providing species‐level identification that</p><p>may inform culture interpretation of pathogens versus contaminants [49]. In the near future,</p><p>it can be also used with adapted software to screen more colonies and to distinguish mono-</p><p>clonal or polyclonal infections due to the same species but potential different clusters [50].</p><p>Lallemand et al. [51] assessed the opportunity to perform MALDI‐TOF identifica-</p><p>tion directly on bone samples for comparison of the different methods available or used</p><p>Table 4.5. Current common NGS platforms used for the diagnosis of clinical microbiology.</p><p>Platforms Manufacturer NSG Generation Technology</p><p>iSeq,MiSeq… Illumina Second Sequencing by synthesis</p><p>Ion Torrent ThermoFisher Second Sequencing by synthesis</p><p>MinION, GridION… Oxford Nanopore Third Exonuclease sequencing and detection</p><p>by electrical conductivity</p><p>Sequel, RSII Pacific</p><p>Biosiences</p><p>Third Sequencing by synthesis</p><p>Figure 4.4. Performance of MALDI‐TOF analysis.</p><p>60 Bone and Joint Infections</p><p>in the routine lab. The use of MALDI‐TOF directly on the specimen provided no useful</p><p>contribution in the routine diagnosis, potentially due to a low bacterial burden in</p><p>operating room samples for BJI diagnosis. The highest sensitivity was obtained with</p><p>optimized culture (85.9%). Direct examination remains insensitive (31.7%) while</p><p>MALDI‐TOF identification represented a 6.3% sensitivity [51]. This method can be</p><p>improved after extraction following enrichment in blood culture bottles, as previously</p><p>evoked. Use of MALDI‐TOF identification on enrichment pellets of bone samples is an</p><p>accurate, rapid, and robust method for bacterial identification of clinical isolates from</p><p>BJI, except for streptococci, whose identification to species level remains difficult as it</p><p>has been previously reported [47].</p><p>On the other hand, Greenwood‐Quaintance et  al. [52] reported a study on the</p><p>assessment of a new technology called PCR‐electrospray ionization mass spectrome-</p><p>try (PCR‐ESI/MS) after the first study by Jacovides et  al. [53] based on a research</p><p>use‐only platform with an old version of the software. In fact, it combined a broad‐</p><p>range PCR with a MALDI‐TOF technology. Obviously, the database is different and</p><p>should be regularly updated. They compared PCR‐ESI/MS results to culture using</p><p>sonicate fluids from 431 explanted‐knee or ‐hip patients. Briefly, one ml of non‐</p><p>concentrated sonicate fluid was thawed and DNA extraction was performed before</p><p>using the PLEX‐ID PCR‐ESI/MS instrument manufactured by Abbott. The sensitivity</p><p>for detecting BJI were 77.6% for PCR‐ESI/MS and 69.7% for culture (p=0.0105) and</p><p>the specificities were 93.5 and 99.3%, respectively (p=0.0002). More sensitive but less</p><p>specific, this method seems to be a useful tool for rapid detection of BJI and can inte-</p><p>grate the arsenal of methods used to reach the causing bacteria. The same group from</p><p>the Mayo clinic published a similar study, using synovial fluid specimens [54]. They</p><p>reported an 81% sensitivity and a 95% specificity for the diagnosis of PJIs. Like with</p><p>other methods, it is necessary to have an appropriate cut‐off for meaningful interpre-</p><p>tation. Other groups have reported the utility of this methodology for detecting DNA</p><p>from hundreds of different microorganisms in whole blood [55]. The detection of the</p><p>microorganism is based on the system’s software and according to two parameters, a</p><p>Q‐score ranging from 0 (low) to 1 (high) representing a relative measure of the strength</p><p>(up to 0.9 is reported) and the level of detection representing a semiquantitative meas-</p><p>ure of the amount of amplified DNA calculated relative to an internal caliper corre-</p><p>sponding to genome equivalents. These parameters should be well controlled in the</p><p>interpretation of the data, in order to avoid misinterpretation of false‐positive or</p><p>false‐negative results. Thus, this method can detect with a level of 15 and a Q‐score of</p><p>0.9 a Micrococcus luteus, but it does not mean infection. Indeed, in Jacovides study</p><p>[53], a high rate of false positivity was highlighted with 49 out of 57 aseptic failure</p><p>patients having positive PCR‐ESI/MS results. Again, like with the NGS, the microor-</p><p>ganism detected, the values found for the interpretation, the clinic, and the multidisci-</p><p>plinary approach remain the keys to success for the diagnosis and the treatment of a</p><p>BJI. Moreover, like with conventional molecular techniques or NGS, microbial DNA</p><p>detection should not be interpreted systematically as infection. DNA persistence in</p><p>clinical samples, bone biopsy, synovial fluid, or tissue following successful treatment</p><p>is real. Bémer et al. [56] reported the detection by 16S PCR of Listeria monocytogenes</p><p>in a perfectly healthy patient, in whom antibiotics were stopped, and who had an</p><p>uneventful follow‐up of one year. With this new technology, one should be aware that</p><p>DNA may persist after successful therapy, and therefore the presence of DNA does</p><p>not always indicate persistent infection [54].</p><p>4 Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques 61</p><p>Key Points (Figure 4.5)</p><p>● Clinical samples should always be obtained for culture in order to detect and test the</p><p>susceptibility of the pathogen(s).</p><p>● The use of non‐culture techniques may increase the rate of etiological diagnosis in</p><p>pediatric BJI, particularly for fastidious microorganisms such as K. kingae.</p><p>● The use of a targeted PCR for S. aureus may shorten the time to diagnosis in cases of</p><p>acute septic arthritis.</p><p>● Non‐culture techniques are a valuable supplemental tool in patients with culture‐</p><p>negative BJI caused by fastidious or slow‐growing microorganisms, allowing earlier</p><p>initiation of pathogen‐adapted therapy.</p><p>● Non‐culture techniques as a supplemental tool improve the detection rate of the path-</p><p>ogen in patients with BJI who have received previous antibiotics.</p><p>References</p><p>1. Trampuz A, Piper KE, Jacobson MJ, et al. Sonication of removed hip and knee prostheses</p><p>for diagnosis of infection. N Engl J Med 2007; 357:654–663.</p><p>2. Berbari EF, Marculescu C, Sia I, et al. Culture‐negative prosthetic joint infection. Clin Infect</p><p>Dis 2007; 45:1113–1119.</p><p>Figure 4.5. Diagnostic algorithm in case of suspected BJI. Conventional culture methods</p><p>should always be performed in order to diagnose and test the susceptibility of the pathogen(s).</p><p>In addition to the conventional methods, non‐culture techniques are useful in special situations,</p><p>as shown in the algorithm.</p><p>62 Bone and Joint Infections</p><p>3. Bellova P, Knop‐Hammad V, Königshausen M, et al. Sonication of retrieved implants improves</p><p>sensitivity in the diagnosis of periprosthetic joint infection. BMC Musculoskelet Disord 2019;</p><p>20: 623.</p><p>4. Baron EJ, Miller JM, Weinstein MP, et al. Executive summary: a guide to utilization of the</p><p>microbiology laboratory for diagnosis of infectious diseases: 2013 recommendations by the</p><p>Infectious Diseases Society of America (IDSA) and the American Society for Microbiology</p><p>(ASM)(a). Clin Infect Dis 2013; 57:485–488.</p><p>5. Corvec S, Portillo ME, Pasticci BM, et al. Epidemiology and new developments in the diagno-</p><p>sis of prosthetic joint infection. Int J Artif Organs 2012; 35:923–934.</p><p>6. Roux AL, Sivadon‐Tardy V, Bauer T, et al. Diagnosis of prosthetic joint infection by beadmill</p><p>processing of a periprosthetic specimen. Clin Microbiol Infect 2011; 17:447–450.</p><p>7. Portillo ME, Salvadó M, Alier A, et al. Advantages of sonication fluid culture for the diagnosis</p><p>of prosthetic joint infection. J Infect 2014; 69:35–41.</p><p>8. Ascione T, Barrack R, Benito N, et  al. General Assembly, Diagnosis, Pathogen Isolation ‐</p><p>Culture Matters: Proceedings of International Consensus on Orthopedic Infections.</p><p>J. Arthroplasty. 2019; 34:S197–S206.</p><p>9. Corvec S, Portillo ME, Vossen JA, et al. Diagnostics. In: AOTrauma Principles of Orthopedic</p><p>Infection Management. 2016: 91–119.</p><p>10. Costerton JW, Post JC, Ehrlich GD, et al. New methods for the detection of orthopedic and</p><p>other biofilm infections. FEMS Immunol Med Microbiol 2011; 61:133–140.</p><p>11. Gomez E, Cazanave C, Cunningham SA, et al. Prosthetic joint infection diagnosis using broad‐</p><p>range PCR of biofilms dislodged from knee and hip arthroplasty surfaces using sonication.</p><p>J Clin Microbiol 2012; 50:3501–3508.</p><p>12. Portillo ME, Salvadó M, Trampuz A, et al. Sonication versus vortexing of implants for diag-</p><p>nosis of prosthetic joint infection. J Clin Microbiol 2013; 51:591–594.</p><p>13. Plouzeau C, Bémer P, Valentin AS, et  al. First experience of a multicenter external quality</p><p>assessment of molecular 16S rRNA gene detection in bone and joint infections. J Clin Microbiol</p><p>2015; 53:419–424.</p><p>14. Reller LB, Weinstein MP, Petti CA. Detection and identification of microorganisms by gene</p><p>amplification and sequencing. Clin Infect Dis 2007; 44:1108–1114.</p><p>15. Fenollar F, Roux V, Stein A, et al. Analysis of 525 samples to determine the usefulness of PCR</p><p>amplification and sequencing of the 16S</p><p>rRNA gene for diagnosis of bone and joint infections.</p><p>J Clin Microbiol 2006; 44:1018–1028.</p><p>16. Kommedal Ø, Kvello K, Skjåstad R, et al. Direct 16S rRNA gene sequencing from clinical</p><p>specimens, with special focus on polybacterial samples and interpretation of mixed DNA chro-</p><p>matograms. J Clin Microbiol 2009; 47:3562–3568.</p><p>17. Bemer P, Plouzeau C, Tande D, et al. Evaluation of 16S rDNA PCR sensitivity and specificity</p><p>for diagnosis of prosthetic‐joint infection: a prospective multicenter cross‐sectional study.</p><p>J Clin Microbiol 2014; 52:3583–3589.</p><p>18. Jun Y, Jianghua L. Diagnosis of periprosthetic joint infection using polymerase chain reaction:</p><p>an updated systematic review and meta‐analysis. Surg Infect (Larchmt) 2018; 19:555–565.</p><p>19. Hartley JC, Harris KA. Molecular techniques for diagnosing prosthetic joint infections – search</p><p>results – PubMed. J Antimicrob Chemother 2014; 69:121–124.</p><p>20. Van Belkum A, Rochas O. Laboratory‐based and point‐of‐care testing for MSSA/MRSA</p><p>detection in the age of whole genome sequencing. Front. Microbiol. 2018; 9:1437.</p><p>21. Ghebremedhin B, Layer F, Konig W, et  al. Genetic classification and distinguishing of</p><p>Staphylococcus species based on different partial gap, 16S rRNA, hsp60, rpoB, sodA, and tuf</p><p>gene sequences. J Clin Microbiol 2008; 46:1019–1025.</p><p>22. El Houmami N, Bzdreng J, Durand GA, et al. Molecular tests that target the RTX locus do not</p><p>distinguish between Kingella kingae and the recently described Kingella negevensis Species.</p><p>J Clin Microbiol 2017; 55:3113–3122.</p><p>23. Ross JJ. Septic Arthritis of Native Joints. Infect. Dis. Clin. North Am. 2017; 31:203–218.</p><p>4 Identification of Pathogens in Bone and Joint Infections by Non‐Culture Techniques 63</p><p>24. Berbari EF, Kanj SS, Kowalski TJ, et al. Executive Summary: 2015 Infectious Diseases Society</p><p>of America (IDSA) clinical practice guidelines for the diagnosis and treatment of native verte-</p><p>bral osteomyelitis in adults. Clin Infect Dis 2015; 61:859–863.</p><p>25. Liu K, Fu J, Yu B, et  al. Meta‐analysis of sonication prosthetic fluid PCR for diagnosing</p><p>periprosthetic joint infection. PLoS One 2018; 13.</p><p>26. Portillo ME, Salvado M, Sorli L, et al. Multiplex PCR of sonication fluid accurately differenti-</p><p>ates between prosthetic joint infection and aseptic failure. J Infect 2012; 65:541–548.</p><p>27. Malandain D, Bémer P, Leroy AG, et al. Assessment of the automated multiplex‐PCR Unyvero</p><p>i60 ITI® cartridge system to diagnose prosthetic joint infection: a multicentre study. Clin</p><p>Microbiol Infect 2018; 24:83.e1–83.</p><p>28. Vasoo S, Cunningham SA, Greenwood‐Quaintance KE, et al. Evaluation of the FilmArray</p><p>blood culture ID panel on biofilms dislodged from explanted arthroplasties for prosthetic joint</p><p>infection diagnosis. J. Clin. Microbiol. 2015; 53:2790–2792.</p><p>29. Dekker JP. Metagenomics for clinical infectious disease diagnostics steps closer to reality.</p><p>J. Clin. Microbiol. 2018; 56:9.</p><p>30. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et al. Impact of contaminating DNA in</p><p>whole‐genome amplification kits used for metagenomic shotgun sequencing for infection diag-</p><p>nosis. J Clin Microbiol 2017; 55:1789–1801.</p><p>31. Street TL, Sanderson ND, Atkins BL, et al. Molecular diagnosis of orthopedic‐device‐related</p><p>infection directly from sonication fluid by metagenomic sequencing. J Clin Microbiol 2017;</p><p>55:2334–2347.</p><p>32. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et al. Identification of prosthetic joint</p><p>infection pathogens using a shotgun metagenomics approach ‐ PubMed. Clin Infect Dis 2018;</p><p>67:1333–1338.</p><p>33. Thoendel M, Jeraldo P, Greenwood‐Quaintance KE, et al. A novel prosthetic joint infection</p><p>pathogen, Mycoplasma Salivarium, identified by metagenomic shotgun sequencing – PubMed.</p><p>Clin Infect Dis 2017; 65:332–335.</p><p>34. Tarabichi M, Alvand A, Shohat N,et al. Diagnosis of Streptococcus canis periprosthetic joint</p><p>infection: the utility of next‐generation sequencing. Arthroplast Today 2018; 4:20–23.</p><p>35. Pham TT, Lazarevic V, Gaia N, et  al. Second periprosthetic joint infection caused by</p><p>Streptococcus dysgalactiae: how genomic sequencing can help defining the best therapeutic</p><p>strategy. Front Med 2020; 7:53.</p><p>36. Huang Z, Zhang C, Li W, et al. Metagenomic next‐generation sequencing contribution in iden-</p><p>tifying prosthetic joint infection due to Parvimonas micra: a case report. J Bone Jt Infect 2019;</p><p>4:50–55.</p><p>37. Sharma H, Ong MR, Ready D, et al. Real‐time whole genome sequencing to control a Streptococcus</p><p>pyogenes outbreak at a national orthopaedic hospital. J Hosp Infect 2019; 103:21–26.</p><p>38. Ivy MI, Thoendel MJ, Jeraldo PR, et al. Direct detection and identification of prosthetic joint</p><p>infection pathogens in synovial fluid by metagenomic shotgun sequencing. J Clin Microbiol</p><p>2018; 56:9.</p><p>39. Sanderson ND, Street TL, Foster D, et al. Real‐time analysis of nanopore‐based metagenomic</p><p>sequencing from infected orthopaedic devices. Biological Sciences Genetics. Microbiology.</p><p>BMC Genomics 2018; 19:714.</p><p>40. Chen MF, Chang CH, Chiang‐Ni C, et al. Rapid analysis of bacterial composition in pros-</p><p>thetic joint infection by 16s rRnA metagenomic sequencing. Bone Jt Res 2019; 8:367–377.</p><p>41. Li M, Zeng Y, Wu Y,et al. Performance of Sequencing Assays in Diagnosis of Prosthetic Joint</p><p>Infection: A Systematic Review and Meta‐Analysis. J. Arthroplasty. 2019; 34:1514–1522.</p><p>42. Torchia MT, Austin DC, Kunkel ST, et al. Next‐generation sequencing vs culture‐based meth-</p><p>ods for diagnosing periprosthetic joint infection after total knee arthroplasty: a cost‐effective-</p><p>ness analysis. J Arthroplasty 2019; 34:1333–1341.</p><p>43. Wang C, Huang Z, Li W, et al. Can metagenomic next‐generation sequencing identify the patho-</p><p>gens responsible for culture‐negative prosthetic joint infection? BMC Infect Dis 2020; 20:253.</p><p>64 Bone and Joint Infections</p><p>44. Deroche, Bémer, Valentin, et al. The right time to safely re‐evaluate empirical antimicrobial</p><p>treatment of hip or knee prosthetic joint infections. J Clin Med 2019; 8:2113.</p><p>45. Carbonnelle E, Beretti JL, Cottyn S, et al. Rapid identification of staphylococci isolated in</p><p>clinical microbiology laboratories by matrix‐assisted laser desorption ionization‐time of flight</p><p>mass spectrometry. J Clin Microbiol 2007; 45:2156–2161.</p><p>46. Kaleta EJ, Clark AE, Cherkaoui A, et al. Comparative analysis of PCR ‐ Electrospray ioniza-</p><p>tion/mass spectrometry (MS) and MALDI‐TOF/MS for the identification of bacteria and</p><p>yeast from positive blood culture bottles. Clin Chem 2011; 57:1057–1067.</p><p>47. Thomin J, Aubin GG, Foubert F, et al. Assessment of four protocols for rapid bacterial identi-</p><p>fication from positive blood culture pellets by matrix‐assisted laser desorption ionization‐time</p><p>of flight mass spectrometry (Vitek® MS). J Microbiol Methods 2015; 115:54–56.</p><p>48. Harris LG, El‐Bouri K, Johnston S, et al. Rapid identification of staphylococci from prosthetic</p><p>joint infections using MALDI‐TOF mass‐spectrometry. Int J Artif Organs 2010; 33:568–574.</p><p>49. Peel TN, Cole NC, Dylla BL, et al. Matrix‐assisted laser desorption ionization time of flight</p><p>mass spectrometry and diagnostic testing for prosthetic joint infection in the clinical microbiol-</p><p>ogy laboratory. Diagn Microbiol Infect Dis 2014; 81: 163–168.</p><p>50. Nagy E, Urbán E, Becker S, et al. MALDI‐TOF MS fingerprinting facilitates rapid discrimina-</p><p>tion of phylotypes I, II and III of Propionibacterium acnes. Anaerobe 2013; 20:20–26.</p><p>51. Lallemand E, Coiffier G, Arvieux C, et al. MALDI‐TOF MS performance compared to direct</p><p>examination, culture, and 16S rDNA PCR for the rapid diagnosis of bone and joint infections.</p><p>Eur J Clin Microbiol Infect Dis 2016; 35:857–866.</p><p>52. Greenwood‐Quaintance KE, Uhl JR, Hanssen AD, et al. Diagnosis of prosthetic joint infection</p><p>by use of PCR‐electrospray ionization mass spectrometry. J Clin Microbiol 2014; 52:642–649.</p><p>53. Jacovides CL, Kreft R, Adeli B, et al. Successful identification of pathogens by polymerase</p><p>chain reaction (PCR)‐based electron spray ionization time‐of‐flight mass spectrometry (ESI‐</p><p>TOF‐MS)</p><p>in culture‐negative periprosthetic joint infection. J Bone Jt Surg ‐ Ser A 2012;</p><p>94:2247–2254.</p><p>54. Melendez DP, Uhl JR, Greenwood‐Quaintance KE, et al. Detection of prosthetic joint infec-</p><p>tion by use of PCR‐electrospray ionization mass spectrometry applied to synovial fluid. J Clin</p><p>Microbiol 2014; 52:2202–2205.</p><p>55. Strålin K, Rothman RE, Özenci V, et al. Performance of PCR/electrospray ionization‐mass</p><p>spectrometry on whole blood for detection of bloodstream microorganisms in patients with</p><p>suspected sepsis. J Clin Microbiol 2020.</p><p>56. Bémer P, Plouzeau C, Tande D, et al. Evaluation of 16S rRNA gene PCR sensitivity and speci-</p><p>ficity for diagnosis of prosthetic joint infection: A prospective multicenter cross‐sectional study.</p><p>J Clin Microbiol 2014; 52:3583–3589.</p><p>65</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>Chapter 5</p><p>History of Bacteriophage Use in Human Infections</p><p>Bacteriophages  –  or literally “bacteria‐eaters,” from the Greek phagein meaning “to</p><p>devour” – were first mentioned in 1915 by Frederick Twort. They owe their therapeutic</p><p>applications to Félix d’Hérelle, who isolated them in 1917 in stool samples of patients</p><p>suffering from shigellosis [1]. Shortly thereafter, d’Herelle used bacteriophages to treat</p><p>bacillary dysentery (shigellosis). This was probably the first attempt of phage application</p><p>to treat pathogenic bacterial infections. The bacteriophage preparation was first ingested</p><p>by d’Herelle and some colleagues in order to evaluate its safety before being administered</p><p>to a 12‐year‐old boy with severe dysentery. After a single application of the anti‐ dysentery</p><p>bacteriophage, the patient’s symptoms terminated, and the boy fully recovered within</p><p>days [2]. Inspired by these results, d’Herelle continued studies on the therapeutic use of</p><p>bacteriophages, carried many non‐randomized trials in humans [3], and co‐founded with</p><p>George Eliava an institute, known today as the “Eliava Institute of Bacteriophages,</p><p>Microbiology and Virology” in Tbilisi, Georgia, to carry out basic bacteriophage research</p><p>and provide bacteriophages to treat human bacterial infections.</p><p>The development of bacteriophages as antimicrobials continued for about three dec-</p><p>ades, i.e. from about 1915–1942 [4]. During this period, bacteriophages were used, among</p><p>other indications, in France against avian typhoid caused by Salmonella gallinarum, and</p><p>in the United States against chronic furunculosis. Phage therapy was also used during the</p><p>Winter War between the former Soviet Union and Finland (1939–1940), with 6,000</p><p>Soviet soldiers treated against streptococcal or staphylococcal wound infections, which</p><p>prevented limb amputations and reduced mortality due to gangrene. Companies such as</p><p>Behring in Germany and Eli Lilly in the United States produced phage preparations</p><p>against streptococci, staphylococci, and Escherichia coli. During World War II in Africa,</p><p>the German army and the allied forces applied bacteriophages against dysentery [5].</p><p>Bacteriophages for Treatment</p><p>of Biofilm Infections</p><p>Mercedes Gonzalez‐Moreno, Paula Morovic, Tamta Tkhilaishvili,</p><p>and Andrej Trampuz</p><p>66 Bone and Joint Infections</p><p>The progress made on the synthesis of penicillin during the 1940s ushered in the era of</p><p>antibiotic use, a golden age of medicine that largely continues to this day, which led to an</p><p>almost complete abandonment of interest in the development of phages as clinically used</p><p>antibacterial agents, especially in the Western countries [4]. In the Eastern countries,</p><p>however, phage therapy was never abandoned, persisting to this day in countries such as</p><p>Poland, Georgia, and Russia. Much knowledge of phage therapy in human patients</p><p>comes from numerous publications in either Russian or Polish journals from that period,</p><p>for example involving the use of oral phages to treat gastrointestinal infections, including</p><p>shigellosis and salmonellosis [4].</p><p>Phage therapy was rediscovered by the English‐language literature starting with the</p><p>work of Smith and Huggins in the 1980s, and progressively gained attention during the</p><p>1990s, followed by the start of human experiments in the 2000s [6]. The first placebo‐</p><p>controlled phase I trial in the United States was published in 2009, and showed no safety</p><p>concerns [7].</p><p>The emergence of multidrug‐resistant bacterial infections has led to recent efforts</p><p>investigating and promoting phage therapy to treat a multitude of infections. Despite the</p><p>costly and time‐consuming requirements for the production of bacteriophages under</p><p>current guidelines in the United States and the European Union, some countries are try-</p><p>ing to accelerate the implementation of phage therapy through the so‐called “Magistral</p><p>Approach.” Belgium, for instance, is currently implementing a pragmatic framework on</p><p>phage therapy that centers on magistral preparation of individual therapeutic bacterio-</p><p>phages by pharmacies, and although the final products will not fully comply with the</p><p>European requirements for medicinal products for human use (Directive 2001/83/EC),</p><p>such magistral phage preparations can be used to treat patients in Belgium [8].</p><p>Principles of Bacteriophage Therapy</p><p>Molecular Background</p><p>Bacteriophages (also referred as phages) are viruses that specifically infect bacteria.</p><p>Phages are the most abundant organisms on Earth, with an estimation of about 1031 par-</p><p>ticles distributed over all ecosystems on our planet. A phage is usually conformed by its</p><p>genome (single or double‐stranded DNA or RNA) encapsulated in a protein capsid,</p><p>which is sometimes completed with a tail and more or less complex appendages (e.g.</p><p>spikes, tail fibers, etc.) (Figure 5.1) [9, 10]. As nonliving microorganisms, they rely on the</p><p>bacterial cellular machinery to reproduce. The viral infection begins by attachment of the</p><p>phage to its bacterial host through specific recognition of one or more receptors on</p><p>the bacterial cell. These receptors can be found in the cell wall, bacterial capsules, slime</p><p>layers, pili, or flagella, often consisting of proteins, lipopolysaccharides, teichoic acids,</p><p>and other cell surface structures serving as irreversible phage‐binding receptors [11].</p><p>Upon recognition of the cell receptors, the phage injects its genetic material into the cyto-</p><p>plasm of the infected cell, and depending on its nature (virulent or temperate), it follows</p><p>the lytic or lysogenic cycle.</p><p>Virulent phages follow the lytic cycle, where the host’s genome is first degraded and the</p><p>bacterial metabolic machinery is employed to copy the viral genome and produce viral</p><p>proteins. After that, the viral particles are self‐assembled, and the bacterial cell is lysed by</p><p>5 Bacteriophages for Treatment of Biofilm Infections 67</p><p>(a) Myoviridae</p><p>(b) Siphoviridae (c) Podoviridae</p><p>Tail</p><p>Capsid</p><p>Tail tube</p><p>Tail</p><p>Long tail fibre</p><p>Central tail fibre or spike</p><p>Capsid</p><p>Capsid</p><p>Tail</p><p>Tail fibre</p><p>Tail fibre</p><p>Central tail</p><p>fibre or spike</p><p>Central tail fibre or spike</p><p>Collar</p><p>Whiskers</p><p>Tail tube and sheath</p><p>Short tail fibre</p><p>Baseplate</p><p>Baseplate</p><p>Figure 5.1. Representative structures of tailed phages. All tailed phages have a capsid that</p><p>encloses and protects the genome and connects to the tail. (a) Phages in the Myoviridae family are</p><p>the only tailed phages with a contractile tail sheath. (b) Both phages belonging to the Myoviridae</p><p>and Siphoviridae families have a baseplate at the distal end of the tail to which receptor‐binding</p><p>proteins (RBPs), such as tail fibers and tail spikes, are attached. (c) Because members of the</p><p>Podoviridae have no baseplate, the RBPs directly attach to the tail. Siphoviridae and Podoviridae</p><p>additionally have a central tail fiber or spike that protrudes from the distal end of the tail or</p><p>baseplate. Reprinted with permission from Nobrega et al. [10].</p><p>68 Bone and Joint Infections</p><p>phage enzymes, releasing the progeny phages and killing the bacterial host.</p><p>Temperate</p><p>phages, on the other hand, can follow the lysogenic cycle. They become latent by inserting</p><p>their genome either as a free plasmid inside the host cell or integrated into the bacterial</p><p>chromosome. By this mean, they propagate to the next generations of bacterial cells.</p><p>Under specific stressful environmental conditions, temperate phages can eventually shift</p><p>towards the lytic cycle. As a result, the phage genome will be excised from the host</p><p>chromosome, replicated, encapsulated, and then the phage particles will be released from</p><p>the host bacterium by cell lysis, causing the death of the bacterial host cell [12]. Temperate</p><p>phages have been described to transfer new genes to their hosts, including antibiotic</p><p>resistance genes. In addition, they can alter the expression of host genes or provide</p><p>protection to the host against infection by other phages [13]. Thus, strictly virulent</p><p>phages, with immediate bactericidal effect, are favored for use in clinical practice.</p><p>Resistance Mechanisms</p><p>Bacteria can develop resistance to phages at any stage of the phage infection cycle, clas-</p><p>sified as adaptive or non‐adaptive mechanism [11]. Common non‐specific anti‐phage</p><p>resistance mechanisms that have been described until now include: (a) preclusion of</p><p>phage adsorption and prevention of nucleic acid entry by surface modifications and</p><p>receptor mutations; (b) superinfection exclusion systems preventing a secondary viral</p><p>infection with the same or a closely related virus; (c) restriction‐modification systems</p><p>responsible for the cleavage of exogenous dsDNA and protection of bacterial genetic</p><p>material; and (d) abortive infection leading to cell death or stasis when phage replication</p><p>takes place. A second line of defense (adaptive defense) is associated with restriction</p><p>enzymes and the CRISPR/Cas system (Clustered Regularly Interspaced Short</p><p>Palindromic Repeats and associated proteins), which can identify and cleave phage</p><p>nucleic acids in a highly specific and effective manner [14].</p><p>Phages can evolve and develop counterstrategies to circumvent bacterial anti‐phage</p><p>mechanisms. Based on their genomic plasticity and rapid replication rates, phages can</p><p>overcome adsorption inhibition by point mutations in specific genes or escape from</p><p>restriction‐modification mechanisms by genome rearrangements. Furthermore, phages</p><p>can use anti‐CRISPR proteins to evade the CRISPR/Cas system, or avoid the abortive</p><p>infection by hijacking bacterial antitoxins [11].</p><p>Contrary to antibiotics, phages have minimal influence on the normal microbiome due</p><p>to their high bacterial species or strain specificity, and have the ability to increase in num-</p><p>ber at the site of infection due to their “multiplicity” [15], which theoretically would</p><p>imply that a little phage dose is sufficient for effective treatment.</p><p>Antibiotic–Bacteriophage Interactions</p><p>Although the use of phages alone would potentially demonstrate clinical success, their</p><p>combination with antibiotics have shown to be more effective than phage monotherapy in</p><p>numerous in vitro and animal studies. These studies have proved statistically or clinically</p><p>significant phage‐antibiotic synergism, biofilm minimization, or reductions in resistance</p><p>emergence [12]. Regardless of the antibiotic resistance state of the bacteria, the combinato-</p><p>rial approach using phages and antibiotics has demonstrated a range of benefits [16]. For</p><p>instance, it has been shown for some phage/antibiotic combinations that sub‐inhibitory</p><p>5 Bacteriophages for Treatment of Biofilm Infections 69</p><p>concentration of antibiotics can foster phage productivity and consequently decrease</p><p>bacterial counts. A restoration of antibiotic sensitivity, by loss of bacterial fitness or by</p><p>phage interaction with the bacterial drug efflux systems, has also been described. Phages</p><p>can additionally work as adjuvants of antibiotics against biofilms by enabling antibiotics to</p><p>reach bacterial cells deeper within biofilms through degradation of the exopolysaccharide</p><p>matrix by depolymerases and by infecting antibiotic‐tolerant persisters (further informa-</p><p>tion in section “Activity of Phages against Bacterial Biofilms and Persisters”). However,</p><p>antagonistic effects can also occur with phage/antibiotic combinations depending on the</p><p>treatment conditions (e.g., dosage, order of administration, timing, etc.). Thus, com-</p><p>binatorial therapies require a careful choice of dosing and time points at which either</p><p>antimicrobial substance is administered. In vitro studies determined better outcomes if</p><p>phage was administered before antibiotics than if antibiotics were introduced before or</p><p>simultaneously with phage. This is possibly due to the killing of host bacteria, which are</p><p>essential for phage production, by antibiotics [12]. Other competing dynamics between</p><p>phage and antibiotics may also play a role. Antibiotics are expected to interfere with aspects</p><p>of bacterial physiology that can be crucial to phage activities, as for instance by interfering</p><p>with bacterial ribosome functioning necessary for phage protein production [17].</p><p>The pharmacokinetic/pharmacodynamic (PK/PD) modeling techniques traditionally</p><p>used for antibiotics differ for phages. The phage concentration is expected to increase at</p><p>the site of infection through their replication in living bacteria. PK/PD models for phage</p><p>therapy should integrate the classical antimicrobial pharmacological view (drug impact</p><p>on the body, drug interactions, absorption, distribution, metabolism, secretion, etc.) with</p><p>the self‐replication characteristic of phages [18].</p><p>The immune system plays also a key role in phage inactivation and/or clearance from</p><p>the body, which may pose a problem maintaining sufficient phage titers for therapeutic</p><p>activity. Based on limited reports on immune response in clinical studies using virulent</p><p>phages, their immunogenicity does not seem to represent a safety risk. Major concerns</p><p>encompass an increase in pro‐inflammatory cytokines as response to the potentially mas-</p><p>sive liberation of bacterial endotoxins after bacterial lysis, as has been observed with the</p><p>use of certain antibiotics [19]. So far, there is not enough evidence‐based data for a better</p><p>understanding of the phage pharmacokinetics and the phage immune interaction as well</p><p>as the clinical relevance of all these parameters.</p><p>Activity of Bacteriophages against Bacterial Biofilms and Persisters</p><p>Biofilms are complex clusters of bacteria, formed by single or multiple species, merged by</p><p>extracellular polymeric substances (EPS) and adhered to surfaces, including living tissue</p><p>or medical devices, among others. Biofilm microorganisms are metabolically less active</p><p>and have a minimal growth rate. Therefore, they are tolerant to many antibiotics [20][21].</p><p>Bacteriophages showed promising results for biofilm eradication due to their multiplicity</p><p>at the infection site, but also by producing specific enzymes that allow them to actively</p><p>penetrate and disrupt biofilms, and for their ability to infect persisters, which are less</p><p>metabolically active bacteria [20].</p><p>Phages encoding EPS‐degrading enzymes are particularly useful against biofilms.</p><p>A diverse group of phage‐encoded enzymes, called depolymerases, capable of degrading</p><p>polymers – either associated with the cell surface (e.g. capsule polysaccharides) to facilitate</p><p>phage adsorption, or EPS involved in biofilm matrix in order to promote phage diffusion</p><p>70 Bone and Joint Infections</p><p>through the biofilm – have been described [22]. Depolymerases can be associated with</p><p>virions, forming part of the phage particle (e.g. in their tail spikes), or be in soluble form.</p><p>Depolymerases derived from phages have been tested against biofilms of different bac-</p><p>terial species, exhibiting dose‐dependent activity and reducing significantly the biofilm</p><p>biomass [20, 23, 24]. Similar to the host specificity of bacteriophages, phage‐associated</p><p>depolymerases can be highly specific for host‐derived EPS. Since different species of bac-</p><p>teria</p><p>Hospitals</p><p>Oxford, UK</p><p>Willem‐Jan Metsemakers, MD</p><p>Department of Trauma Surgery</p><p>University Hospitals Leuven</p><p>Department of Development and</p><p>Regeneration, KU Leuven</p><p>Leuven, Belgium</p><p>Beat K. Moor, MD</p><p>Service of Traumatology and</p><p>Orthopedic Surgery</p><p>Wallis Hospital Center</p><p>Martigny, Switzerland</p><p>xiv List of Contributors</p><p>Pilar Morata, PhD</p><p>Biochemistry and Molecular Biology</p><p>Faculty of Medicine</p><p>University of Malaga</p><p>Malaga, Spain</p><p>T. Fintan Moriarty, PhD</p><p>AO Research Institute Davos</p><p>AO Foundation</p><p>Davos, Switzerland</p><p>Mario Morgenstern, MD</p><p>Centre for Musculoskeletal Infections</p><p>University Hospital Basel</p><p>Department of Orthopaedic and Trauma</p><p>Surgery</p><p>University of Basel</p><p>Basel, Switzerland</p><p>Paula Morovic</p><p>Charité‐Universitätsmedizin</p><p>Corporate Member of Freie Universität</p><p>Berlin</p><p>Humboldt‐ Universität zu Berlin, and</p><p>Berlin Institute of Health</p><p>Center for Musculoskeletal Surgery</p><p>Berlin, Germany</p><p>Andreas Marc Müller, MD</p><p>Centre for Musculoskeletal Infections</p><p>University Hospital Basel</p><p>Department of Orthopaedic and Trauma</p><p>Surgery</p><p>University of Basel</p><p>Basel, Switzerland</p><p>Roger L. Nation, PhD</p><p>Drug Delivery</p><p>Disposition and Dynamics</p><p>Monash Institute of Pharmaceutical Sciences</p><p>Monash University (Parkville Campus)</p><p>Melbourne, Australia</p><p>Maria Eugenia Portillo, MD, PhD</p><p>Clinical Microbiology</p><p>Complejo Hospitalario de Navarra</p><p>Irunlarrea</p><p>Pamplona, Navarra, Spain</p><p>Nora Renz, MD</p><p>Charité‐Universitätsmedizin</p><p>Corporate member of Freie Universität</p><p>Berlin</p><p>Humboldt‐ Universität zu Berlin, and</p><p>Berlin Institute of Health</p><p>Center for Musculoskeletal Surgery</p><p>Berlin, Germany</p><p>R. Geoff Richards, PhD</p><p>AO Research Institute Davos</p><p>AO Foundation</p><p>Davos, Switzerland</p><p>Olivier Robineau, MD</p><p>Infectious Diseases Department</p><p>University Hospitals of Tourcoing</p><p>Lille University</p><p>Tourcoing, France</p><p>Arick P. Sabin, MD</p><p>Department of Internal Medicine</p><p>Gundersen Health System</p><p>La Crosse, WI, USA</p><p>Parham Sendi, MD</p><p>Centre for Musculoskeletal Infections</p><p>University Hospital Basel</p><p>Institute for Infectious Diseases</p><p>University of Bern</p><p>Bern, Switzerland</p><p>Eric Senneville, MD, PhD</p><p>Infectious Diseases Department</p><p>University Hospitals of Tourcoing</p><p>Lille University</p><p>Tourcoing, France</p><p>Fritz Sörgel, PhD</p><p>IBMP‐Institute for Biomedical and</p><p>Pharmaceutical Research</p><p>Nürnberg‐Heroldsberg, and</p><p>Institute of Pharmacology</p><p>Faculty of Medicine</p><p>University of Duisburg‐Essen</p><p>Essen, Germany</p><p>List of Contributors xv</p><p>Christoph Spormann, MD</p><p>Upper Extremities</p><p>Hirslanden Clinic and Endoclinic Zürich</p><p>Zürich, Switzerland</p><p>Priya Sukhtankar</p><p>NIHR Southampton Clinical Research</p><p>Facility</p><p>University Hospital Southampton</p><p>NHS Foundation Trust</p><p>Southampton, UK</p><p>Tamta Tkhilaishvili, MD</p><p>Charité‐Universitätsmedizin</p><p>Corporate member of Freie Universität</p><p>Berlin</p><p>Humboldt‐ Universität zu Berlin, and</p><p>Berlin Institute of Health</p><p>Center for Musculoskeletal Surgery</p><p>Berlin, Germany</p><p>Andrej Trampuz, MD</p><p>Charité‐Universitätsmedizin</p><p>Corporate member of Freie Universität</p><p>Berlin</p><p>Humboldt‐ Universität zu Berlin, and</p><p>Berlin Institute of Health</p><p>Center for Musculoskeletal Surgery</p><p>Berlin, Germany</p><p>Rihard Trebse, MD</p><p>Valdoltra Orthopaedic Hospital Ankaran</p><p>University of Ljubljana Medical Faculty</p><p>Ljubljana, Slovenija</p><p>Ilker Uçkay, MD</p><p>Department of Orthopedic Surgery</p><p>Infectiology</p><p>Balgrist University Hospital</p><p>University of Zürich</p><p>Zürich, Switzerland</p><p>Felix W.A. Waibel, MD</p><p>Department of Orthopedic Surgery</p><p>Balgrist University Hospital</p><p>University of Zürich</p><p>Zürich, Switzerland</p><p>Pablo Yagupsky, MD</p><p>Clinical Microbiology Laboratory</p><p>Soroka University Medical Center</p><p>Ben Gurion University of the Negev</p><p>Beer Sheva, Israel</p><p>Werner Zimmerli, MD</p><p>Infectious Diseases</p><p>Interdisciplinary Unit for Orthopedic</p><p>Infections</p><p>Kantonsspital Baselland</p><p>University of Basel</p><p>Liestal, Switzerland</p><p>Matthias A. Zumstein, MD</p><p>Shoulder, Elbow, and Orthopedic Sports</p><p>Medicine</p><p>Orthopaedics Sonnenhof, Bern</p><p>Bern, Switzerland</p><p>xvi</p><p>The knowledge in the field of bone and joint infections is steadily increasing. Looking at</p><p>PubMed® shows that the number of publications in the field of septic arthritis, peripros-</p><p>thetic joint infection, and osteomyelitis has been strongly rising for 8–15 years. Thus, it</p><p>seems natural that a new edition of this textbook dealing with bone and joint infections</p><p>would facilitate the daily clinical decisions of specialized and non‐specialized physicians.</p><p>Due to an unstoppable specialization of infectious diseases, a textbook with information</p><p>weighted by dedicated specialists in the field is still useful, despite the unlimited availabil-</p><p>ity of information in the internet. The main drawback of unfiltered publications in medi-</p><p>cal databases is the necessity of the appropriate selection by the clinician. For this</p><p>purpose, specialized clinical experience is needed. Therefore, a textbook written by recog-</p><p>nized experts in their field allows to make the correct clinical decisions.</p><p>In the 2nd edition, all chapters have been updated to the current scientific knowledge. In</p><p>addition, four new chapters have been added. Since the correct diagnosis is the base of the</p><p>rational treatment, a chapter with diagnostic algorithms for the most important topics in</p><p>the field of bone and joint infections is added. Implants increase the susceptibility to infec-</p><p>tion by compromising the local host defense. Therefore, literally every microorganism may</p><p>cause periprosthetic joint infection. A new chapter presents clinical information about</p><p>such rare infections caused by unusual bacteria. In routine diagnostic labs, many microbi-</p><p>ologists may lack specialized knowledge in the field of non‐culture techniques. Therefore,</p><p>a new chapter presents these methods for the application in patients with bone and joint</p><p>infection. The correct use of these methods is discussed. This allows the rational use of</p><p>non‐culture techniques without wasting laboratory capacity and money. Finally, the</p><p>fourth new chapter is dedicated to bacteriophages, which have been recently used in the</p><p>treatment of different chronic infections also in western countries. This therapeutic tech-</p><p>nique is of increasing interest, due to the rising problem of multiresistant bacteria. Up to</p><p>now, controlled clinical trials have been missing. The presentation of the current knowl-</p><p>edge may motivate clinicians to plan such studies in patients with biofilm infections.</p><p>I hope that the new edition, written by a dedicated multidisciplinary team of specialists, is</p><p>a useful guide to the rational management of patients with bone and joint infections.</p><p>Werner Zimmerli, MD</p><p>Preface to the Second Edition</p><p>xvii</p><p>Correct and rapid diagnosis and treatment of bone and joint infections require the</p><p>collaboration of many different specialists. In the field of bone and joint infections, there</p><p>is almost no evidence based on controlled trials. Therefore, learning from the clinical</p><p>experience of experts is particularly important for the management of such infections.</p><p>Up to now, a comprehensive, internationally available textbook dealing with infections of</p><p>the locomotor system has been missing. The present book fills this gap.</p><p>An editor of a book on bone and joint infections must have a broad knowledge of the</p><p>pertinent literature. In addition, he or she should know the different competent special-</p><p>ists in the field. When John Wiley & Sons asked Werner Zimmerli to edit this book, they</p><p>could not have come up with a better choice. He is a competent specialist in the field of</p><p>infections of the locomotor system and also has contacts with a network of international</p><p>specialists.</p><p>Werner Zimmerli started his professional career in infectious diseases in Geneva, in the</p><p>group of Francis A. Waldvogel, a well‐known specialist of osteomyelitis, whose pioneer</p><p>publications are still widely cited. Under his leadership, Werner Zimmerli started his studies</p><p>on implant‐related infections. In the early 1980s, it was rather an exception for an infectious</p><p>disease specialist to invade a field for which orthopedic surgeons were responsible.</p><p>Together with Daniel Lew and Pierre Vaudaux, he was able to define host factors that</p><p>are responsible</p><p>produce different EPS components, depolymerase active against the polysaccha-</p><p>rides produced by one species may not act on that produced by other bacteria [24].</p><p>However, some depolymerases are capable of degrading EPS of several genera [23].</p><p>Moreover, some bacteriophages can induce their host bacteria to produce and release</p><p>depolymerases, which could be a phage mechanism to make the biofilm matrix more</p><p>porous, facilitating infection by progeny bacteriophage or, alternatively, a fight response</p><p>by infected bacteria, seeking to facilitate movement away from the focus of infection [24],</p><p>leading in any case to a disaggregation of the biofilm.</p><p>Unlike other antimicrobials, bacteriophages replicate within their host cells, which can</p><p>result in self‐sustaining infections with ongoing amplification leading to an increasing</p><p>number of bacteriophages. The localized spread of phage progeny continues infecting</p><p>and killing more bacteria, which is called multiplicity at the infection site. These mecha-</p><p>nisms require a critical mass of host bacteria at the same location, which is typically the</p><p>case in biofilm infections [24]. Hence, by spreading through the biofilm, bacteriophages</p><p>can progressively remove the biofilm and reduce the potential for regrowth.</p><p>The regrowth of bacteria within the biofilm is thought to arise from the presence of per-</p><p>sisters. Unlike resistant bacterial cells, where resistance mechanisms are based on genetic</p><p>changes that block antimicrobial activity, persisters present a transient non‐heritable phe-</p><p>notype that is thought to be less sensitive to antibiotics because the cells are not undergoing</p><p>cellular activities that antibiotics can corrupt, which results in tolerance [25]. Thus, persist-</p><p>ers can remain viable over the course of antibiotic exposure and repopulate the biofilm</p><p>when the levels of antibiotic drop, causing the relapse of the infection. Some studies have</p><p>reported on the ability of bacteriophages to infect persisters, and initiate a productive lytic</p><p>infection when persisters switch to normal growth, ultimately causing their lysis [24].</p><p>New concepts are emerging nowadays in the design of phage‐based treatments to max-</p><p>imize phage therapy efficacy and minimize the likelihood of resistance emergence. A</p><p>schematic illustration of phage‐based treatments for biofilm removal is shown in</p><p>Figure 5.2 [20].</p><p>Phage-based treatments for biofilm removal</p><p>Phage therapy Phage-derived enzymesGenetically modified phages Phage + Antibiotics</p><p>Single phages Phage cocktails</p><p>Figure 5.2. Main phage‐based treatments for biofilm removal. Reprinted with permission from</p><p>Ferriol‐González and Domingo‐Calap [20].</p><p>5 Bacteriophages for Treatment of Biofilm Infections 71</p><p>Designing phage cocktails that include phages against multiple species has been shown</p><p>to be especially effective against multi‐species biofilms [23, 26]. Phage cocktails, besides</p><p>conferring activity against a broader host range, can help also in preventing the emer-</p><p>gence of phage‐resistant bacteria if multiple phages active against a given target are</p><p>included in the cocktail [20]. However, in order to avoid possible undesired effects when</p><p>using phage cocktails, a rational approach to designing cocktails is crucial. In a phage</p><p>cocktail, the various phages should not compete with each other, in order to minimize the</p><p>risk for reduction of efficacy. In addition, the mechanisms of phage resistance by bacteria</p><p>should be different in order to minimize the risk of cross‐resistance [26].</p><p>Bacteriophages can be genetically modified to improve their bacterial killing proper-</p><p>ties. Existing examples of genetically engineered phages include a phage with altered tail</p><p>fiber proteins to extend its host range [23], a phage designed to produce a soluble hydro-</p><p>lase that enhances biofilm degradation, a temperate phage turned into a lytic phage by</p><p>removal of all genes related to lysogeny or a chimeric phage encoding a short peptide</p><p>with broad‐spectrum anti‐biofilm effect [20].</p><p>Bacteriophage Susceptibility Testing: the Phagogram</p><p>Phage therapy is still not approved in most parts of the world, and extensive research on</p><p>its efficacy and safety is still to be conducted. However, if the particular conditions accord-</p><p>ing to the Declaration of Helsinki (article 37) are met and bacteriophages are to be applied,</p><p>the magistral approach (“compounded” drug product in United States) is to be followed.</p><p>In practice, it means that, only when no other option of treatment is available, a phage</p><p>preparation is prescribed by a physician for an individual patient and prepared by the</p><p>hospital pharmacist following strict safety regulations [8]. Due to the high host‐specificity</p><p>of bacteriophages (mostly infecting single species or even single strains of bacteria), when</p><p>preparing the phage solution for an individual patient, it is important to select bacterio-</p><p>phages active against the patient’s isolated strain. To this end, similarly to an antibiogram</p><p>(antibiotic susceptibility testing), a so‐called “phagogram” needs to be performed [8].</p><p>Various methods for testing bacteria susceptibility to bacteriophages have been</p><p>described, such as the spot test, efficacy of plating (EOP), or killing assays. The simplest</p><p>method among them is the spot test, where small droplets of a bacteriophage lysate are</p><p>applied on a plate prepared with the bacterial strain to be tested and the appearance of a</p><p>clear zone (lysis) determines bacterial susceptibility to the bacteriophage. However, lysis</p><p>observed by this method may be the result not only from the bacteriophage infection that</p><p>gives rise to lysis and production of new phage, but also due to residues of bacteriocins on</p><p>the phage lysate that kills bacteria, or to the phages themselves causing abortive infections</p><p>or lysis from without, leading to false‐positive results [27]. Performing an EOP assay, by</p><p>plating different titers of bacteriophages, quantifying plaque‐forming units (PFU), and</p><p>comparing it to the PFU count of a reference bacterial strain can provide more informa-</p><p>tion on the efficacy of a particular bacteriophage [27]. Nevertheless, the absence of plaque</p><p>formation does not necessarily correlate with a lack of bacteriophage ability for a produc-</p><p>tive infection. Plaque formation might depend on several factors, including phage diffu-</p><p>sion in agar, adsorption rate, electrolyte requirements, growth phase of the host, etc. [28].</p><p>Killing assays, in which bacteria and bacteriophages are incubated together in liquid</p><p>medium and the optical density or heat flow production are measured as indicators of</p><p>bacterial presence, represent useful methods for determination of the minimum phage</p><p>72 Bone and Joint Infections</p><p>titer needed for successful bacteria killing or to better monitor phage virulence [29, 30]. On</p><p>the other hand, these assays are usually less cost‐efficient than clinical laboratory tests due</p><p>to a higher instrumental cost, minimum automation, or limitations in their throughput.</p><p>A standardized method that is easy to perform, fast, and available to everyone is still to be</p><p>developed. Currently, several projects for the development of an automated, reliable, and</p><p>reproducible phagogram technique are ongoing. Some examples include the PHAGOGramme</p><p>project under development by Pherecydes Pharma (https://www.pherecydes‐pharma.</p><p>com/phagogramme.html), the combined PhageBank™and HRQT™approaches of the</p><p>clinical‐stage company Adaptive Phage Therapeutics (http://www.aphage.com/the‐</p><p>science/#phagebank) or the PhagoFlow project as a joint effort of different institutions</p><p>including the Charité‐Universitätsmedizin, the Bundeswehr Hospital Berlin, the Leibniz</p><p>Institute DSMZ, and the Fraunhofer ITEM (https://www.phagoflow.de/en/phagogram/).</p><p>Experimental and Clinical Evidence with Bacteriophage Treatment</p><p>Despite the long history of use of phages for antibacterial therapy since their discovery</p><p>in the early 1900s, and even with the availability of phage products for the treatment of</p><p>bacterial</p><p>infections in some countries (e.g. Georgia, Poland, Russia), extensive in vitro</p><p>and experimental studies as well as clinical trials to fulfill the requirements for phage</p><p>therapy according to good manufacturing practice guidelines are lacking [31].</p><p>The efficacy of phage therapy has been investigated for bloodstream, gastrointestinal,</p><p>urinary tract, and respiratory infections, and burn wounds [19, 26]. We limit our focus on</p><p>experimental and clinical evidence of phage therapy in bone and joint infections.</p><p>Most preclinical studies investigated the efficacy of bacteriophages on monomicrobial</p><p>S. aureus or P. aeruginosa infections demonstrating large reduction of planktonic bacte-</p><p>ria, successful prevention of bacterial adherence to foreign material, and synergism</p><p>between antibiotics and phages to eradicate biofilms [32]. However, numerous experi-</p><p>mental limitations need to be addressed. For instance, limited data exist on phages active</p><p>against S. epidermidis despite its high prevalence in implant‐associated infections, its</p><p>strong biofilm‐forming capability, and its extensive resistance to antibiotics [33].</p><p>Moreover, in vivo models that replicate the joint and peri‐implant microenvironment are</p><p>lacking, which makes the translation of preclinical findings into clinical settings difficult</p><p>[34]. One promising in vivo model published by Carli et al. in 2017 [35] replicates accu-</p><p>rately the clinical setting of total joint replacement, and could therefore be adopted in the</p><p>future for phage therapy testing.</p><p>Other studies reported a concentration dependency of phage therapy, suggesting that</p><p>low‐titer phage administration or single instead of multiple doses are unlikely to be suc-</p><p>cessful. In addition, considering, for instance, possible vascular impairments in open</p><p>fractures or the wish for reduction of systemic effects, local treatment is often preferred</p><p>in bone and joint infections [36]. Hence, in aiming for phage stability and appropriate</p><p>release kinetics during treatment, an important part of research in phage therapy is</p><p>focused on the encapsulation of phages into sustained release systems. Numerous strate-</p><p>gies regarding bacteriophage formulation and encapsulation are being implemented,</p><p>showing promising outcomes under experimental settings [36]. Still, great challenges</p><p>involve a rational design of carriers loaded with precise doses of encapsulated phage able</p><p>to support controlled releases in patients.</p><p>Osteomyelitis is another clinical field where phage therapy has been applied, often</p><p>using anti‐staphylococcal bacteriophages. A summary of clinical studies on phage</p><p>https://www.pherecydes-pharma.com/phagogramme.html</p><p>https://www.pherecydes-pharma.com/phagogramme.html</p><p>http://www.aphage.com/the-science/#phagebank</p><p>http://www.aphage.com/the-science/#phagebank</p><p>https://www.phagoflow.de/en/phagogram/</p><p>5 Bacteriophages for Treatment of Biofilm Infections 73</p><p>therapy for musculoskeletal infections is presented in Table 5.1 [32]. The largest clinical</p><p>study with 120 participants was conducted in Tbilisi, Georgia, assessing the therapeutic</p><p>efficacy of a custom‐made staphylococcal cocktail against arthritis and osteomyelitis.</p><p>The summarized results do not allow us to evaluate phage efficacy. All 120 patients had</p><p>complete recovery of osteomyelitis and/or arthritis, namely 9 patients with phage therapy</p><p>alone, 51 patients with phage plus antibiotics, and 60 patients with antibiotics alone.</p><p>These results are much better in each treatment group than could be expected.</p><p>In western countries, due to strict regulations in application of phage therapy, clinical</p><p>experience with bacteriophages is limited to individual cases with a total of 5 case reports</p><p>and 1 case series published between 2017 and 2019. As shown in Table  5.1, two case</p><p>reports investigated the use of bacteriophages in osteomyelitis, another two in prosthetic</p><p>Table 5.1. Human clinical studies on phage therapy for musculoskeletal infections. Reprinted</p><p>with  permission from Onsea et al. [32].</p><p>Reference</p><p>Sample</p><p>size</p><p>Patient</p><p>characteristics Intervention Outcome</p><p>Lang et al.,</p><p>1979</p><p>7 PJI (n = 2), OM</p><p>(n = 1), Septic</p><p>arthritis (n = 1),</p><p>Spinal infection</p><p>(n = 1), FRI (n = 2).</p><p>Phages adapted to isolated strains.</p><p>Administration either topical or by</p><p>injection through a draining system.</p><p>Some cases received combination</p><p>treatment with antibiotics.</p><p>5/7 treated</p><p>Recurrence of spinal</p><p>infection and one FRI.</p><p>Kutateladze</p><p>and Adamia,</p><p>2010</p><p>120 Patients with</p><p>staphylococcal</p><p>OM or arthritis.</p><p>Three groups:</p><p>‐ antibiotics (n = 60)</p><p>‐ phage monotherapy (n = 9)</p><p>‐ phage + antibiotics (n = 51).</p><p>Administration of Eliava</p><p>staphylococcal phage preparation</p><p>topically or intravenously.</p><p>100% success rate in</p><p>all groups.</p><p>Slopek et al.,</p><p>1987</p><p>100 Purulent arthritis</p><p>and myositis</p><p>(n = 19), OM of</p><p>the long bones</p><p>(n = 40), FRI</p><p>(n = 41).</p><p>Administration locally and/or orally.</p><p>Some cases received combination</p><p>treatment with antibiotics.</p><p>Success rates:</p><p>‐ purulent arthritis</p><p>and myositis: 89.5%</p><p>‐ OM of the long</p><p>bones: 95%</p><p>‐ FRI: 90.2%</p><p>Weber‐</p><p>Dabrowska</p><p>et al., 2000</p><p>81 OM of the long</p><p>bones (n = 40),</p><p>FRI (n = 41).</p><p>Administration locally and/or orally.</p><p>Unclear if some patients received</p><p>combination treatment with</p><p>antibiotics.</p><p>Success rates:</p><p>OM of the long bones:</p><p>95%FRI 60%</p><p>Vogt et al.,</p><p>2017</p><p>1 OM Repeated dosing of phage cocktail.</p><p>Pyo bacteriophage through draining</p><p>system, in combination with</p><p>antibiotic therapy.</p><p>Eradication of the</p><p>infection.</p><p>Ferry et al.,</p><p>2018a</p><p>1 OM</p><p>(post‐radiation)</p><p>Application of customised phage</p><p>cocktail every 3 d, in combination</p><p>with intravenous antibiotic therapy.</p><p>Patient died 45 d after</p><p>treatment due to</p><p>cancer progression.</p><p>Ferry et al.,</p><p>2018b</p><p>1 PJI Single intraoperative injection of a</p><p>customised phage cocktail in</p><p>combination with intravenous</p><p>antibiotic therapy.</p><p>Eradication of the</p><p>infection.</p><p>(Continued )</p><p>74 Bone and Joint Infections</p><p>joint infection, and one in a fracture‐related infection. The applied bacteriophages were</p><p>used to target P. aeruginosa, S. aureus, Acinetobacter baumannii, Klebsiella pneumoniae,</p><p>S. epidermidis, and E. faecalis, respectively. Bacteriophages were administered either</p><p>intravenously or locally, and in combination with intravenous antibiotics. Eradication of</p><p>infection was seen in all case reports except for one, where the outcome of infection was</p><p>unclear due to death caused by the primary disease of the patient.</p><p>Although bacteriophages have so far demonstrated good efficacy and safety, experi-</p><p>ence is still lacking, and comprehensive and well‐organized studies on the production and</p><p>processing of bacteriophages, their administration, and dosage, as well as exhaustive</p><p>clinical monitoring of results, are still needed.</p><p>Local Delivery and Systemic Bacteriophage Application</p><p>A major hurdle of phage therapy is the achievement of a sufficient number of phages at</p><p>the site of infection to accomplish therapeutic activity. Phage ability to disseminate</p><p>throughout the body strongly depends on the route of administration and the initial</p><p>phage dose [37]. Administration of high or repeated phage doses might increase the</p><p>chances for successful distribution. Furthermore, encapsulation of phages might allow a</p><p>controlled phage release and act as a shield against chemical degradation or immunologi-</p><p>cal neutralization, prolonging its systemic circulation period [18]. Routes of phage appli-</p><p>cation include the use of parenteral administration, being oral dosing, topical, and</p><p>aerosolization also commonly applied. A summary of some advantages and disadvan-</p><p>tages of the administration routes can be seen in Table 5.2 ([19]).</p><p>Systemic Delivery</p><p>Systemic phage delivery by intravenous, intraperitoneal, or intramuscular injection</p><p>allows rapid phage dissemination in different organs and tissues such as the liver, spleen,</p><p>kidneys, and lungs [38]. A nearly complete recovery of administered phages was shown</p><p>Table 5.1. (Continued)</p><p>Reference</p><p>Sample</p><p>size</p><p>Patient</p><p>characteristics</p><p>Intervention Outcome</p><p>Nir‐Paz et al.,</p><p>2019</p><p>1 FRI Intravenous repeated administration</p><p>of customised phage cocktail, in</p><p>combination with intravenous</p><p>antibiotic therapy.</p><p>Eradication of the</p><p>infection (after two</p><p>phage therapy</p><p>regimens).</p><p>Tkhilaishvili</p><p>et al., 2019</p><p>1 PJI Repeated dosing of customised</p><p>phage cocktail, in combination with</p><p>intravenous antibiotic therapy.</p><p>Eradication of the</p><p>infection.</p><p>Onsea et al.,</p><p>2019</p><p>4 OM Repeated dosing of BFC1 cocktail</p><p>or Pyo bacteriophage cocktail in</p><p>combination with intravenous</p><p>antibiotic therapy.</p><p>Eradication of the</p><p>infection in all cases.</p><p>Abbreviations: PJI, periprosthetic joint infection; OM, osteomyelitis; FRI, fracture‐related infection.</p><p>5 Bacteriophages for Treatment of Biofilm Infections 75</p><p>several minutes after intravenous application [37]. Phage distribution was documented in</p><p>the heart, skeletal muscles, bladder, thymus, bone narrow, lungs, and brain but not yet in</p><p>joints, bone, or eyes [18].</p><p>The use of highly purified phage preparations without bacterial components or endo-</p><p>toxins is essential to minimize the risk of side effects due to impurities.</p><p>Oral or Inhaled Delivery</p><p>Oral route of administration has been successfully used in gastrointestinal infections.</p><p>However, phage stability in acidic environments in the stomach and duodenum may</p><p>reduce the phage concentration or activity [38]. Thus, protection of phages from the gas-</p><p>tric acidity could be achieved by phage encapsulation, as shown in a study against</p><p>Salmonella spp. [39].</p><p>In respiratory infections, liquid and dry powder phage formulations were investigated for</p><p>nebulization and inhalation for topical delivery in acute and chronic lung infections [40].</p><p>Table 5.2. Routes of administration for phage therapy. Reprinted with permission from Romero‐Calle</p><p>et al. [19].</p><p>Delivery</p><p>Route Advantages Disadvantages Mitigations to Hurdles</p><p>Intraperitoneal Higher dosage volumes</p><p>possible. Diffusion to</p><p>other sites.</p><p>Extent of diffusion to other</p><p>sites may be overestimated</p><p>in humans (most data from</p><p>small animals).</p><p>Multiple delivery sites.</p><p>Intramuscular Phages delivered at</p><p>infection site.</p><p>Slower diffusion of phages</p><p>(possibly). Lower dosage</p><p>volumes.</p><p>Multi‐dose courses.</p><p>Subcutaneous Localized and systemic</p><p>diffusion.</p><p>Lower dosage volumes. Multi‐dose courses.</p><p>Intravenous Rapid systemic diffusion. Rapid clearing of phages by</p><p>the immune system.</p><p>In vivo selection of</p><p>low‐immunogenic</p><p>phages may be possible.</p><p>Topical High dose of phages</p><p>delivered at infection site.</p><p>Run‐off from target site if</p><p>phages suspended in liquid.</p><p>Incorporate phages into</p><p>gels and dressings.</p><p>Suppository Slow, stable release of</p><p>phages over long time.</p><p>Limited applications/sites.</p><p>Risk of insufficient dosing.</p><p>Technically challenging to</p><p>manufacture.</p><p>Careful consideration of</p><p>phage kinetics required.</p><p>Oral Ease of delivery. Higher</p><p>dosage volumes</p><p>possible.</p><p>Stomach acid reduces</p><p>phage titer. Non‐specific</p><p>adherence of phages to</p><p>stomach contents and other</p><p>microflora.</p><p>Add calcium carbonate</p><p>to buffer pH.</p><p>Microencapsulation to</p><p>deliver phages to target</p><p>area.</p><p>Aerosol Relative ease of delivery.</p><p>Can reach poorly</p><p>perfused regions of</p><p>infected lungs.</p><p>High proportion of phages</p><p>lost. Delivery can be</p><p>impaired by mucus and</p><p>biofilms.</p><p>Use of depolymerases</p><p>to reduce mucus.</p><p>76 Bone and Joint Infections</p><p>Local Delivery</p><p>Effective local delivery of antibacterial substances is essential in patients with biofilm</p><p>infections associated either with implanted medical devices or chronic wounds, since anti-</p><p>bacterial drugs have limited activity in such infections. Bacteriophages may have better</p><p>efficacy in such infections, provided that they reach the infectious site. Therefore, a major</p><p>focus is being set in the implementation of drug carriers to allow local and prolonged</p><p>release of phages. Unfortunately, there is insufficient data available on processing phages</p><p>into well‐defined pharmaceutical formulations, their long‐term stability, and impact on</p><p>phage efficacy in vivo [40]. Wound healing is one of the therapeutic areas where local</p><p>application of phages has received a lot of attention. Advances in the development of</p><p>phage formulations including hydrogels, liposome entrapment, or phage‐immobilized</p><p>wound dressings have led to increasing successful rates in the topical application of phage</p><p>therapy [41]. Numerous reviews report on the widespread clinical use of phage prepara-</p><p>tions for the treatment of skin infections, and purulent and surgical wounds, mostly by</p><p>the former Soviet Union countries. In Europe, the project Phagoburn, launched in June</p><p>2013, was the first prospective multicentric, randomized, single‐blind, and controlled</p><p>clinical trial on phage therapy to treat Escherichia coli and Pseudomonas aeruginosa skin</p><p>infections in burn patients [42]. It allowed significant advances regarding the regulatory</p><p>framework of phage therapy as well.</p><p>Some clinical cases and preclinical studies also support the effective local delivery of</p><p>bacteriophages to treat local bone infections (see section “Experimental and Clinical</p><p>Evidence with Bacteriophage Treatment”). Currently, the application of phages in</p><p>patients with severe musculoskeletal infections has generally consisted in local adminis-</p><p>tration through a draining system (Figure 5.3) [43]. Although this approach has shown</p><p>successful outcomes, it has the drawback of the usage of the drainage tube as delivery</p><p>route, which could favor the emergence of superinfections, besides being a cumbersome</p><p>method. Thus, the optimization of local phage delivery strategies might help in overcom-</p><p>ing these issues. Some examples are an engineered hydrogel for controlled delivery of</p><p>phage targeting Pseudomonas aeruginosa to the site of orthopedic infections [44] or the</p><p>use of fibrin glue for sustained delivery of viable phages [45]. Phages have also been</p><p>immobilized on surfaces for the prevention of biofilm formation with examples on uri-</p><p>nary catheters or on nylon sutures for wound healing applications [36].</p><p>Outlook and Future Perspectives</p><p>The rising threat of multiresistant bacterial infections has brought together many research</p><p>institutions, hospitals, and the industry in a joint effort to seek alternative treatments to</p><p>the conventional use of antibiotics.</p><p>Phages have unique features that make them convincing antibacterial agents, alone or</p><p>in combination with other antimicrobials, while the constraints associated with the imple-</p><p>mentation of phage therapy could be overcome through a combination of proper phage</p><p>selection, effective formulation, and greater clinician understanding of and familiarity</p><p>with product application.</p><p>Phages have been used to treat bacterial infections since their discovery, being for dec-</p><p>ades and also today the standard of care in several countries of Eastern Europe and hav-</p><p>ing demonstrated clinical success in recent compassionate care cases in Western Europe</p><p>and the United States, with no serious adverse events reported to date.</p><p>5 Bacteriophages for Treatment of Biofilm Infections 77</p><p>The increasing number of publications that have appeared during the last decade and</p><p>the growing interest of the industry in phage therapy represent very encouraging progress</p><p>in addressing the knowledge gap required for phage therapeutic applications.</p><p>Key Points</p><p>● With increasing antimicrobial resistance of bacteria, there is a rising interest in the</p><p>therapeutic potency of bacteriophages.</p><p>(a) (b)</p><p>(c) (d)</p><p>Figure 5.3. Phage therapy of pelvic osteomyelitis (pathogen: pan‐resistant Pseudomonas</p><p>aeruginosa). (a) Removal of foreign bodies and surgical debridement of necrotic tissue. (b)</p><p>Preparation of a wound filler impregnated with the phage solution. (c) Insertion of the instilla-</p><p>tion tubes before wound closure. (d) Daily administration of 50 ml of the bacteriophage suspen-</p><p>sion after pre‐treatment of the wound with bicarbonate buffer (over one week). Reprinted with</p><p>permission from Vogt et al. [43].</p><p>78 Bone and Joint Infections</p><p>● Biofilm infections are tolerant to most antibiotics. Phage therapy of such infections</p><p>could be an attractive new option.</p><p>● There are a few clinical studies showing that bacteriophages are able to eradicate mus-</p><p>culoskeletal infections without serious adverse events. However, controlled trials of</p><p>high quality are still lacking.</p><p>● Despite current restrictions in the application of phage therapy, commercialization of</p><p>phage‐based technologies is gaining interest in Western countries.</p><p>References</p><p>1. D’Herelle F. On an invisible microbe antagonistic toward dysenteric bacilli: brief note by Mr.</p><p>F. D’Herelle, presented by Mr. Roux. 1917. Res Microbiol, 2007. 158(7):553–554.</p><p>2. Sulakvelidze A, Alavidze Z, Morris JG, Jr. Bacteriophage therapy. Antimicrob Agents</p><p>Chemother, 2001. 45(3):649–659.</p><p>3. Dublanchet A and Bourne S. The epic of phage therapy. Can J Infect Dis Med Microbiol, 2007.</p><p>18(1):15–18.</p><p>4. Green S, Ma L, Maresso A. (2019). Phage therapy, in T.M. Schmidt (ed.) Encyclopedia of</p><p>Microbiology (Fourth Edition), pp. 485–495 (Oxford: Academic Press).</p><p>5. Moelling K, Broecker F, Willy C. A wake‐up call: We need phage therapy now. Viruses, 2018.</p><p>10(12). doi: 10.3390/v10120688.</p><p>6. Wittebole X, De Roock S, Opal SM. A historical overview of bacteriophage therapy as an</p><p>alternative to antibiotics for the treatment of bacterial pathogens. Virulence, 2014.</p><p>5(1):226–235.</p><p>7. Rhoads DD, Wolcott RD, Kuskowski MA, et al. Bacteriophage therapy of venous leg ulcers in</p><p>humans: results of a phase I safety trial. J Wound Care, 2009. 18(6):237–243.</p><p>8. Pirnay JP, Verbeken G, Ceyssens PJ, et  al. The magistral phage. Viruses, 2018. 10(2). doi:</p><p>10.3390/v10020064.</p><p>9. Dion MB, Oechslin F, Moineau S. Phage diversity, genomics and phylogeny. Nat Rev Microbiol,</p><p>2020. 18(3):125–138.</p><p>10. Nobrega FL, Vlot M, de Jonge PA, et  al. Targeting mechanisms of tailed bacteriophages.</p><p>Nat Rev Microbiol, 2018. 16(12):760–773.</p><p>11. Orzechowska B and Mohammed M. (2019). The war between bacteria and bacteriophages, in</p><p>M. Mishra (ed.) Growing and Handling of Bacterial Cultures, (London: IntechOpen). doi:</p><p>10.5772/intechopen.87247.</p><p>12. Morrisette T, Kebriaei R, Lev KL, et al. Bacteriophage therapeutics: A primer for clinicians on</p><p>phage‐antibiotic combinations. Pharmacotherapy, 2020. 40(2):153–168.</p><p>13. Howard‐Varona C, Hargreaves KR, Abedon ST, et al. Lysogeny in nature: mechanisms, impact</p><p>and ecology of temperate phages. ISME J, 2017. 11(7):1511–1520.</p><p>14. Ofir G and Sorek R. Contemporary phage biology: from classic models to new insights. Cell,</p><p>2018. 172(6):1260–1270.</p><p>15. Loc‐Carrillo C and Abedon ST. Pros and cons of phage therapy. Bacteriophage, 2011.</p><p>1(2):111–114.</p><p>16. Tagliaferri TL, Jansen M, Horz HP. Fighting pathogenic bacteria on two fronts: Phages and</p><p>antibiotics as combined strategy. Front Cell Infect Microbiol, 2019. 9(22). doi: 10.3389/</p><p>fcimb.2019.00022.</p><p>17. Abedon ST. Phage‐antibiotic combination treatments: Antagonistic impacts of antibiotics on the</p><p>pharmacodynamics of phage therapy? Antibiotics, 2019. 8(4). doi: 10.3390/antibiotics8040182.</p><p>18. Dąbrowska K. Phage therapy: What factors shape phage pharmacokinetics and bioavailabil-</p><p>ity? Systematic and critical review. Med Res Rev, 2019. 39(5):2000–2025.</p><p>5 Bacteriophages for Treatment of Biofilm Infections 79</p><p>19. Romero‐Calle D, Guimarães‐Benevides R, Góes‐Neto A, et al. Bacteriophages as alternatives</p><p>to antibiotics in clinical care. Antibiotics, 2019. 8(3). doi: 10.3390/antibiotics8030138.</p><p>20. Ferriol‐González C and Domingo‐Calap P. Phages for biofilm removal. Antibiotics, 2020. 9(5).</p><p>doi: 10.3390/antibiotics9050268.</p><p>21. Stewart PS. Antimicrobial tolerance in biofilms. Microbiol Spectr, 2015. 3(3). doi: 10.1128/</p><p>microbiolspec.MB‐0010‐2014.</p><p>22. Pires DP, Oliveira H, Melo LDR, et al. Bacteriophage‐encoded depolymerases: their diversity</p><p>and biotechnological applications. Appl Microbiol Biotechnol, 2016. 100(5):2141–2151.</p><p>23. Geredew‐Kifelew L, Mitchell JG, Speck P. Mini‐review: efficacy of lytic bacteriophages on</p><p>multispecies biofilms. Biofouling, 2019. 35(4):472–481.</p><p>24. Harper DR, Parracho HMRT, Walker J, et al. Bacteriophages and biofilms. Antibiotics, 2014.</p><p>3(3):270–284.</p><p>25. Balaban NQ, Helaine S, Lewis K, et al. Definitions and guidelines for research on antibiotic</p><p>persistence. Nat Rev Microbiol, 2019. 17(7):441–448.</p><p>26. Kortright KE, Chan BK, Koff JL, et al. Phage therapy: A renewed approach to combat antibi-</p><p>otic‐resistant bacteria. Cell Host Microbe, 2019. 25(2):219–232.</p><p>27. Khan‐Mirzaei M and Nilsson AS. Isolation of phages for phage therapy: A comparison of</p><p>spot tests and efficiency of plating analyses for determination of host range and efficacy. PLoS</p><p>One, 2015. 10(3). doi: 10.1371/journal.pone.0118557.</p><p>28. Abedon S. (2018). Detection of bacteriophages: phage plaques, in D. Harper, et  al. (ed.)</p><p>Bacteriophages, pp. 1–32. (Basel: Springer, Cham).</p><p>29. Xie Y, Wahab L, Gill JJ. Development and validation of a microtiter plate‐based assay for</p><p>determination of bacteriophage host range and virulence. Viruses, 2018. 10(4). doi: 10.3390/</p><p>v10040189.</p><p>30. Tkhilaishvili T, Di Luca M, Abbandonato G, et  al. Real‐time assessment of bacteriophage</p><p>T3‐derived antimicrobial activity against planktonic and biofilm‐embedded Escherichia coli by</p><p>isothermal microcalorimetry. Res Microbiol, 2018. 169(9):515–521.</p><p>31. Alavidze Z, Ami‐nov R, Betts A, et al. Silk route to the acceptance and re‐implementation of</p><p>bacteriophage therapy. Biotechnol J, 2016. 11(5):595–600.</p><p>32. Onsea J, Wagemans J, Pirnay JP, et al. Bacteriophage therapy as a treatment strategy for ortho-</p><p>paedic‐device‐related infections: where do we stand? Eur Cell Mater, 2020. 39:193–210.</p><p>33. Akanda ZZ, Taha M, Abdelbary H. Current review—The rise of bacteriophage as a unique thera-</p><p>peutic platform in treating peri‐prosthetic joint infections. J Orthop Res, 2018. 36(4):1051–1060.</p><p>34. Jie K, Deng P, Cao H, et al. Prosthesis design of animal models of periprosthetic joint infection</p><p>following total knee arthroplasty: A systematic review. PLoS One, 2019. 14(10). doi: 10.1371/</p><p>journal.pone.0223402.</p><p>35. Carli AV, Bhimani S, Yang X, et al. Quantification of peri‐implant bacterial load and in vivo</p><p>biofilm formation in an innovative, clinically representative mouse model of periprosthetic</p><p>joint infection. J Bone Joint Surg Am, 2017. 99(6). doi: 10.2106/jbjs.16.00815.</p><p>36. Malik DJ, Sokolov IJ, Vinner GK, et al. Formulation, stabilisation and encapsulation of bac-</p><p>teriophage for phage therapy. Adv Colloid Interface Sci, 2017. 249:100–133.</p><p>37. Huh H, Wong S, St. Jean J, et al. Bacteriophage interactions with mammalian tissue: Therapeutic</p><p>applications. Adv Drug Deliv Rev, 2019. 145:4–17.</p><p>38. Ryan EM, Gorman SP, Donnelly RF, et al. Recent advances in bacteriophage therapy: how</p><p>delivery routes, formulation, concentration and timing influence the success of phage therapy.</p><p>J Pharm Pharmacol, 2011. 63(10):1253–1264.</p><p>39. Colom J, Cano‐Sarabia M, Otero J, et al. Liposome‐encapsulated bacteriophages for enhanced</p><p>oral phage therapy against Salmonella spp. Appl Environ Microbiol, 2015. 81(14):4841–4849.</p><p>40. Chang RYK, Wallin M, Lin Y, et al. Phage therapy for respiratory infections. Adv Drug Deliv</p><p>Rev, 2018. 133:76–86.</p><p>41. Chang RYK, Morales S, Okamoto Y, et al. Topical application of bacteriophages for treatment</p><p>of wound infections. Transl Res, 2020. 220:153–166.</p><p>80 Bone and Joint Infections</p><p>42. Jault P, Leclerc T, Jennes S, et al. Efficacy and tolerability of a cocktail of bacteriophages to</p><p>treat burn wounds infected by Pseudomonas aeruginosa (PhagoBurn): a randomised, con-</p><p>trolled, double‐blind phase 1/2 trial. Lancet Infect Dis, 2019. 19(1):35–45.</p><p>43. Vogt D, Sperling S, Tkhilaishvili T, et al. Beyond antibiotic therapy“ – Zukünftige antiinfek-</p><p>tiöse Strategien – Update 2017. Der Unfallchirurg, 2017. 120(7):573–584.</p><p>44. Wroe JA, Johnson CT, García AJ. Bacteriophage delivering hydrogels reduce biofilm formation</p><p>in vitro and infection in vivo. J Biomed Mater Res A,</p><p>2020. 108(1):39–49.</p><p>45. Rubalskii E, Ruemke S, Salmoukas C, et al. Fibrin glue as a local druge‐delivery system for</p><p>bacteriophage PA5. Sci Rep, 2019. 9(1). doi: 10.1038/s41598‐018‐38318‐4.</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>81</p><p>Chapter 6</p><p>Chronic osteomyelitis requires prolonged antibiotic treatment, has a high recurrence rate,</p><p>and can cause irreversible damage. Also, the number of orthopedic device‐related infec-</p><p>tions is projected to increase [1,2]. Therefore, adequate antibiotic treatment and surgical</p><p>prophylaxis are critical. Therapeutic success is primarily determined by the antimicrobial</p><p>activity against the infecting pathogen and the rate and extent of antibiotic penetration</p><p>into bone. Adequate bone penetration has to be ensured as antibiotics need to reach</p><p>effective concentrations at the infection site to kill bacterial pathogens. Therefore, studying</p><p>the time‐course and extent of bone penetration before launching a clinical effectiveness</p><p>trial is important. The aim of this chapter is to review the pharmacokinetics (PK) and</p><p>pharmacodynamics (PD) of antibiotics in bone and present methods that support optimized</p><p>evidence‐based selection of antibiotic dosage regimens.</p><p>Pharmacokinetics</p><p>The time‐course and magnitude of drug concentrations in the body and particularly at</p><p>the site of action determine the drug effects. Therefore, it is important to study PK, which</p><p>describes the relationship between the dose of a drug and the resulting time‐course of</p><p>drug concentrations at various spaces in the body  [3,4]. PK processes include drug</p><p>absorption from the site of administration into the systemic circulation (except if admin-</p><p>istered directly into the bloodstream), distribution from the systemic circulation into tis-</p><p>sues, and elimination via metabolism, renal excretion, or both. Most frequently PK is</p><p>characterized based on drug concentrations measured in plasma or serum. However, in</p><p>treating bone infections, adequate antibiotic concentrations must be achieved at the site of</p><p>infection in bone. Numerous clinical studies have been conducted to quantify antibiotic</p><p>concentrations in bone.</p><p>Pharmacokinetics and Pharmacodynamics</p><p>of Antibiotics in Bone</p><p>Cornelia B. Landersdorfer, Jürgen B. Bulitta, Roger L. Nation, and Fritz Sörgel</p><p>82 Bone and Joint Infections</p><p>Bone is a heterogeneous tissue, where the organic bone matrix represents 30–35% of</p><p>total bone mass and includes collagen fibrils (~90%), glycoproteins, proteoglycans, and</p><p>extracellular fluid. Blood vessels in bone are located in Haversian and Volkmann’s canals</p><p>that transverse the bone matrix. Bone cells represent only 1–2% of total bone mass and</p><p>in their most mature form as osteocytes are trapped inside the bone matrix. The inorganic</p><p>matrix (65–70%) consists of calcium phosphate crystals (hydroxyapatite) deposited inside</p><p>the organic matrix. Due to this heterogeneous composition, most likely neither bacteria</p><p>nor antibiotics distribute evenly throughout the bone tissue.</p><p>The site of the pathogens in bone is not well known. Based on their size (e.g., ~1 μm for</p><p>Staphylococcus aureus), bacteria are expected to distribute through the Haversian and</p><p>Volkmann’s canals (~70 μm diameter) in bone, but not into the hydroxyapatite crystals. S.</p><p>aureus can enter into and survive in osteoblasts, which may explain relapses. In addition,</p><p>this pathogen adheres to components of the bone matrix such as collagen [5,6].</p><p>Techniques to separate the many different components of bone and measure concen-</p><p>trations in each are lacking. Therefore, the vast majority of published studies are based</p><p>on homogenized bone samples, and the total drug concentrations in bone homogenate</p><p>are reported. For the interpretation of bone penetration results, it is important to note</p><p>that only free (unbound) drug is microbiologically active. However, total drug concentra-</p><p>tions in bone homogenate, provided they are reliably determined and analyzed by popu-</p><p>lation PK modeling and Monte Carlo simulations, may be more predictive of therapeutic</p><p>success than serum concentrations, as the latter are further removed from events in the</p><p>bone matrix.</p><p>Bone Sample Preparation and Analysis</p><p>In contrast to plasma or serum, there is no specific guidance available for drug analysis in</p><p>bone or other tissues. However, validated accurate and reproducible sample preparation</p><p>and drug determination procedures are undoubtedly critical. When interpreting the</p><p>results of published studies, it is important to consider the analytical techniques used.</p><p>After bone resection, adhering blood and soft tissue is often removed from the sample.</p><p>Excess blood due to intraoperative soaking can result in biased results, for example arti-</p><p>ficially high bone concentrations for a drug with low bone penetration but high blood</p><p>concentrations. Samples are usually separated into cancellous bone (the inner part of the</p><p>long bones) and cortical bone. Cancellous bone has a higher degree of vascularization, a</p><p>higher percentage of extravascular fluid, and a lower percentage of inorganic matrix than</p><p>cortical bone, which can cause differences in antibiotic penetration.</p><p>For efficient extraction of the antibiotic, bone samples need to be homogenized. When</p><p>bone samples are pulverized under liquid nitrogen in a cryogenic mill, this provides a very</p><p>fine powder; this procedure is highly reproducible and is applicable to thermally unstable</p><p>drugs (e.g., β‐lactam antibiotics) that are prone to degradation during grinding without</p><p>freezing. Therefore, this method is preferable to slicing, grinding by mortar and pestle, or</p><p>using mixers without cooling, as was often applied in earlier studies before more recent</p><p>technology was developed. During drug extraction from the homogenized sample, suffi-</p><p>cient and reproducible recovery and stability of the drug need to be ensured.</p><p>Calibration standards and quality control samples are necessary for accurate and pre-</p><p>cise drug determination and should be prepared in drug‐free bone powder instead of</p><p>plasma, serum, or buffer. An internal calibration standard should be added to each</p><p>6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 83</p><p>sample to improve the analytical accuracy and precision. Older studies frequently deter-</p><p>mined drug concentrations by microbiological assay. Newer studies have mainly employed</p><p>high‐performance liquid chromatography (HPLC) or liquid chromatography–tandem</p><p>mass spectrometry (LC–MS/MS), offering improved sensitivity and specificity. These</p><p>chromatographic procedures have been shown to be generally superior to bioassays when</p><p>analyzing bone samples [5]. Bone penetration studies should report details on the chosen</p><p>methods for sample preparation and analysis, and the recovery, bias, and precision.</p><p>Concentrations in bone homogenate are typically reported as mg/kg of total bone</p><p>mass. Some studies report concentrations in relation to bone volume, organic bone mass,</p><p>or interstitial fluid, or correct for blood content. Potential differences in reporting need</p><p>to be considered when comparing results between studies.</p><p>Pharmacokinetic Sampling and Data Analysis</p><p>In recent years, some studies have used microdialysis in an attempt to define a concentra-</p><p>tion time‐course of unbound antibiotic within bone of individual uninfected subjects;</p><p>some such studies are considered below. However, that technique is not without limita-</p><p>tions and it is not known whether the concentration measured represents the concentra-</p><p>tion that may occur at the site of an infection  [7]. As noted above, most studies have</p><p>involved the collection of bone samples, most commonly from people undergoing ortho-</p><p>pedic surgery. Usually, only one bone sample can be taken per patient, and a blood sam-</p><p>ple is taken at the same time. Most studies report bone penetration as the concentration</p><p>ratio between bone and serum</p><p>or plasma at one time point. However, due to different</p><p>kinetics of drug concentrations in plasma and bone, the concentration ratios change over</p><p>time until eventually an equilibrium has been reached during the terminal phase. This</p><p>phenomenon (system hysteresis) hampers the interpretation of results and comparison</p><p>between drugs and studies when samples are taken at different times post dosing.</p><p>A better measure for the extent of bone penetration is to calculate the area under the</p><p>concentration–time curve (AUC) in bone and compare it to the AUC in plasma or serum.</p><p>This approach considers the full time‐course of the concentration profiles in bone and</p><p>plasma (or serum). Instead of collecting the bone and blood samples at the same time</p><p>point after the dose for all patients, samples should be spread out over a time period to</p><p>support PK modeling. Based on such study designs, investigators have averaged the con-</p><p>centrations at each time point and derived the average AUCs in bone and plasma (naive</p><p>averaging) [8–10]. Alternatively, one PK function has been fit to the concentration–time</p><p>data from all patients (naive pooling) and the AUC integrated [11]. While these approaches</p><p>remove the issue of time‐dependent concentration ratios, they only consider the average</p><p>concentration–time profile and ignore the true biological variability between patients.</p><p>Population PK analysis is the most powerful approach for the analysis of sparse data</p><p>(e.g., one bone sample per patient) and accounts for the average rate and extent of bone</p><p>penetration and interpatient variability [7,12–16]. A recent advance has been the use of a</p><p>physiologically based population PK model to describe measured concentrations of cip-</p><p>rofloxacin in plasma and bone. This type of modeling incorporates consideration of dif-</p><p>ferent compartments and blood flows within bone [7]. By fitting each patient’s data in the</p><p>perspective of the concentrations from the other patients, the most likely concentration–</p><p>time‐course in bone and serum and the AUC can be predicted for each patient. Estimating</p><p>the rate of bone penetration enables recommendations on the administration time of</p><p>84 Bone and Joint Infections</p><p>antibiotic prophylaxis before surgery. An existing population PK model can also be used</p><p>to identify the optimal timing of bone and plasma samples in future bone penetration</p><p>studies.</p><p>Bone penetration is usually studied in joint replacement patients with uninfected bone</p><p>as such patients are more easily recruited than osteomyelitis patients. The condition of</p><p>the bone samples is likely more homogeneous among joint replacement patients than</p><p>patients with various stages and locations of bone infections; therefore, results of differ-</p><p>ent studies can be more readily compared. However, antibiotic concentrations might dif-</p><p>fer between infected and uninfected bone. Reactive hyperemia could increase the blood</p><p>flow into bone, whereas pus or sequesters might limit the distribution of antibiotics into</p><p>bone. To date, few studies have been performed in patients with bone infections, which do</p><p>not enable a systematic comparison of penetration between infected and uninfected bone.</p><p>Presence of ischemic, calcified, or arthritic tissues, bone cysts, or fat in the cancellous</p><p>bone may affect antibiotic distribution. Different types of bone (e.g., hip, knee, sternum)</p><p>and influences on blood circulation, for example tourniquet application or internal mam-</p><p>mary artery harvesting, may also affect antibiotic bone concentrations.</p><p>Penetration of Antibiotics into Bone</p><p>Figure 6.1 presents an overview of the extent of bone penetration by antibiotic or antibi-</p><p>otic group. Each symbol represents the median or average bone‐to‐serum (or plasma)</p><p>concentration ratio from one clinical study, and the lines indicate the median per antibiotic</p><p>3</p><p>2</p><p>1</p><p>0.5</p><p>0.3</p><p>0.2</p><p>0.1</p><p>0.05</p><p>qu</p><p>ino</p><p>lon</p><p>es</p><p>m</p><p>ac</p><p>ro</p><p>lid</p><p>es</p><p>cli</p><p>nd</p><p>am</p><p>yc</p><p>in</p><p>rifa</p><p>m</p><p>pic</p><p>in</p><p>lin</p><p>ez</p><p>oli</p><p>d</p><p>gly</p><p>co</p><p>pe</p><p>pt</p><p>ide</p><p>s</p><p>pe</p><p>nic</p><p>illi</p><p>ns</p><p>P-la</p><p>cta</p><p>m</p><p>as</p><p>e-</p><p>inh</p><p>ibi</p><p>to</p><p>rs</p><p>ce</p><p>ph</p><p>alo</p><p>sp</p><p>or</p><p>ins</p><p>fos</p><p>fom</p><p>yc</p><p>in</p><p>fu</p><p>sid</p><p>ic</p><p>ac</p><p>id</p><p>tet</p><p>ra</p><p>cy</p><p>cli</p><p>ne</p><p>s</p><p>M</p><p>ed</p><p>ia</p><p>n</p><p>bo</p><p>ne</p><p>/s</p><p>er</p><p>um</p><p>co</p><p>nc</p><p>en</p><p>tr</p><p>at</p><p>io</p><p>n</p><p>ra</p><p>tio</p><p>Figure 6.1. Bone penetration for different antibiotic groups [5]. Each symbol represents the</p><p>median or average bone‐to‐serum or bone‐to‐plasma concentration ratio from one clinical study.</p><p>A concentration ratio of 11.6 for levofloxacin from one study is off the scale [14]. The lines</p><p>represent the group medians. A comprehensive reference list for papers up to and including the</p><p>year 2012 can be found in reference [5]; more recent papers are listed at the end of this chapter.</p><p>6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 85</p><p>or antibiotic group. In total, 140 studies (until July 2020) were included. Most concentration</p><p>ratios were reported directly in the published studies; sometimes they were calculated</p><p>from the reported concentrations or read from plots. Tables 6.1 and 6.2 list the range of</p><p>average concentration ratios for various antibiotics; for some studies dispersion around</p><p>the average is provided  [7,8,10–12,14–38]. Unless otherwise indicated, concentration</p><p>ratios are based on total concentrations in bone homogenate and serum (plasma); micro-</p><p>dialysis sampling of bone was used in a few studies. Some studies are discussed in more</p><p>detail in the text.</p><p>Table 6.1. Bone penetration of selected fluoroquinolones and macrolides.</p><p>Antibiotic and bone</p><p>condition</p><p>Range of</p><p>time since</p><p>last dose</p><p>Range of average bone/</p><p>serum concentration</p><p>ratios Bone or surgery type</p><p>Bio‐analytical</p><p>method</p><p>Ciprofloxacin</p><p>uninfected 0.5–13 h 0.27–1.2 Hip, knee, skull, HPLC [10,17,18]</p><p>uninfected 0.5–20 h Increased over time;</p><p>organic bone matrix to</p><p>plasma partition</p><p>coefficients were 3.39 for</p><p>cortical and 5.11 for</p><p>cancellous bone; based</p><p>on physiologically based</p><p>population PK modeling</p><p>Hip replacement,</p><p>fracture of femoral</p><p>neck, humerus,</p><p>femur, tibia, pelvic</p><p>column</p><p>HPLC [7]</p><p>osteomyelitis 2–4.5 h 0.42 debridement surgery HPLC [17]</p><p>Levofloxacin</p><p>uninfected 0.7–2 h 0.36–1.0 Hip, other HPLC [19–21]</p><p>uninfected 1–3 h 3.8 ± 2.1 at 1 h,</p><p>11.6 ± 6.4 at 2 h,</p><p>20.9 ± 11.9 at 3 h; based</p><p>on simulations from</p><p>population PK modeling</p><p>Hip or knee</p><p>replacement; cortical</p><p>bone analyzed</p><p>HPLC [14]</p><p>Ofloxacin</p><p>uninfected 0.5–12 h 0.09–1.04 Hip, nasal bone,</p><p>mastoid process</p><p>HPLC [22–24]</p><p>Moxifloxacin</p><p>uninfected 1.5–5 h 0.33–1.05 Hip, knee, sternum,</p><p>manubrium</p><p>HPLC [21,25–27]</p><p>Azithromycin</p><p>uninfected 0.5–6.5</p><p>days</p><p>2.5–6.3 Alveolar bone Bioassay [28,29]</p><p>Telithromycin</p><p>uninfected 3.3–24 h 1.5–2.6 Ethmoid bone HPLC [8]</p><p>HPLC, high‐performance liquid chromatography.</p><p>Table 6.2. Bone penetration of selected beta‐lactams.</p><p>Antibiotic and bone</p><p>condition</p><p>Range of time since last</p><p>dose</p><p>Range of average bone/serum</p><p>concentration ratios Bone or surgery type Bio‐analytical method</p><p>Amoxicillin</p><p>Uninfected 0.5–6 h 0.03–0.31 Hip, jaw bioassay [11,30,31],</p><p>LC‐MS/MS [32]</p><p>Clavulanic acid</p><p>Uninfected 0.5–6 h 0.01–0.14 Hip bioassay [11,30],</p><p>LC‐MS/MS [32]</p><p>Cefepime</p><p>Uninfected 1–2 h 0.46–0.76a Hip HPLC [33]</p><p>Ceftazidime</p><p>Uninfected 2 h 0.54 Cardiac surgery Bioassay [34]</p><p>Ischemic bone 1–2 h 0.04–0.08 foot HPLC [35,36]</p><p>Cefazolin</p><p>Uninfected 2.5–24 h</p><p>(microdialysis)</p><p>0.74 ± 0.36 L sternum,</p><p>0.99 ± 0.59 R sternum; based on AUCs</p><p>Sternal cancellous bone HPLC [37]</p><p>Ceftriaxone</p><p>Uninfected 1–5 h 0.08 ± 0.04 at 1 h,</p><p>1.12 ± 1.29 at 5 h; based on simulations</p><p>from population PK modeling</p><p>Hip or knee replacement;</p><p>cancellous bone</p><p>HPLC [12]</p><p>Cefuroxime</p><p>Uninfected 0.25–8 h</p><p>(microdialysis)</p><p>short‐term infusion: cancellous bone 1.03,</p><p>cortical bone 0.35</p><p>continuous infusion: cancellous bone 1.15,</p><p>cortical bone 0.65</p><p>based on unbound AUCs after population</p><p>PK modeling</p><p>Knee replacement;</p><p>microdialysis cancellous</p><p>and cortical tibia bone</p><p>HPLC [15]</p><p>Ertapenem</p><p>Uninfected 2–28 h 0.025; from population PK modeling Joint replacement or lower</p><p>limb amputation</p><p>LC‐MS [16]</p><p>Flucloxacillin</p><p>Uninfected 3–71 min prior to skin</p><p>incision</p><p>hip replacement:</p><p>0.078 femoral head;</p><p>0.071 femoral neck</p><p>knee replacement:</p><p>0.056 femur;</p><p>0.055 tibia</p><p>Hip or knee replacement HPLC [38]</p><p>HPLC, high‐performance liquid chromatography; LC‐MS and LC–MS/MS, liquid chromatography–mass spectrometry and liquid chromatography–</p><p>tandem mass spectrometry.</p><p>aAssuming a bone density of 1 kg/L.</p><p>88 Bone and Joint Infections</p><p>The data summarized in Figure 6.1 reveal some systematic differences across antibiotic</p><p>groups, which may be due to different physicochemical and binding characteristics.</p><p>There was an approximate five‐fold difference in the median bone/serum concentra-</p><p>tion ratio of the antibiotic class with the lowest ratio (penicillins) and the class with</p><p>the highest ratio (macrolides). Median bone‐to‐serum concentration ratios were 0.50</p><p>for quinolones and 0.48 for linezolid. Despite large differences in chemical structure,</p><p>clindamycin, rifampicin, glycopeptides, fosfomycin, and fusidic acid had comparable</p><p>median concentration ratios of 0.23–0.35. Penicillins, cephalosporins, and β‐lacta-</p><p>mase inhibitors showed median concentration ratios of 0.16, 0.21, and 0.22. Figure 6.1</p><p>also demonstrates large variability within each antibiotic group. This is most noticea-</p><p>ble for the macrolides for which median ratios spanned a very wide range. Part of this</p><p>variability may relate to the differing physicochemical characteristics across the mac-</p><p>rolides as the ratio for azithromycin was reported to be substantially higher than that</p><p>for spiramycin [5]. Also contributing to the variable ratios reported may be differences</p><p>in analytical procedures (e.g., some studies included in Figure 6.1 used bioassays while</p><p>others used liquid chromatography methods) and, as discussed, collection of samples</p><p>at different times after drug administration.</p><p>Fluoroquinolones</p><p>Fluoroquinolones are frequently used in bone infections, and show one of the highest</p><p>median extents of bone penetration of all antibiotic groups with bone‐to‐serum concen-</p><p>tration ratios mostly between 0.3 and 1.2 (Figure 6.1 and Table 6.1). The high penetration</p><p>may be partly due to binding of quinolones to the calcium in bone. As only free antibiotic</p><p>is considered microbiologically active, the quinolone concentrations available for antimi-</p><p>crobial action are likely lower than the total bone concentrations. The concentration</p><p>ratios of most quinolones tend to increase with time since the last dose, as recently</p><p>reported for ciprofloxacin, indicating slow redistribution from bone back into the blood-</p><p>stream [7]. Quinolones generally penetrate well into cells. This could be advantageous for</p><p>treatment of S. aureus osteomyelitis, since S. aureus was shown to penetrate into and</p><p>survive in osteoblasts in vitro [6].</p><p>Multiple studies in different patient groups have examined the bone penetration of</p><p>ciprofloxacin (Table 6.1). In the most recently reported study, which was conducted in</p><p>uninfected patients undergoing orthopedic surgery (n = 39), the average observed cortical</p><p>bone‐to‐plasma concentration ratio was 0.67 at 0.5 to 2 h and 5.1 at 13 to 20 h; for cancel-</p><p>lous bone the respective average ratios were 0.77 and 4.4 [7]. The novel physiologically</p><p>based population PK modelling in that study estimated the partition coefficient of the</p><p>organic bone matrix was 3.39 for cortical and 5.11 for cancellous bone. Fong et al. [17]</p><p>compared ciprofloxacin concentrations in cortical bone from patients without (n = 18,</p><p>hip or knee replacement or osteotomy) and with (n = 10) osteomyelitis. Concentrations</p><p>in infected bone were 30–100% higher than in uninfected bone. As serum concentrations</p><p>were also higher in osteomyelitis patients, the average bone‐to‐serum concentration ratios</p><p>were approximately 0.4 in both patient groups.</p><p>Bone penetration of levofloxacin was evaluated in four studies (Table 6.1). In patients</p><p>undergoing bone surgery (n = 9) or decubitus ulcer debridement (n = 12), the bone‐to‐</p><p>serum concentration ratios were 0.36 for cortical (n = 6) and 0.85 ± 0.40 for cancellous</p><p>(n = 14) bone [19]. In 12 hip replacement patients, ratios of 1.0 ± 0.4 for cortical and</p><p>6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 89</p><p>0.5 ± 0.1 for cancellous bone were reported [20]. Concentration ratios of 0.42 ± 0.04 in</p><p>cortical and 0.54 ± 0.05 in cancellous bone were found in eight hip replacement patients [21].</p><p>In a more recent study, substantially higher cortical bone‐to‐plasma concentration ratios</p><p>were reported, the explanation for which remains to be determined [14].</p><p>Moxifloxacin was studied in four studies, which showed consistently high penetration</p><p>considering the range of different bone types (Table 6.1) and methods for sample homog-</p><p>enization (hand mincing, sonication, cryogenic mill). In all moxifloxacin studies, the pen-</p><p>etration into cancellous and cortical bone was similar. Utilizing a cryogenic mill and</p><p>population PK analysis, the bone‐to‐serum AUC ratios in 24 hip replacement patients</p><p>were 0.80 (10th–90th percentile for between‐patient variability: 0.51–1.26) for cortical</p><p>and 0.78 (0.42–1.44) for cancellous bone [25].</p><p>Macrolides and Telithromycin</p><p>Studies on macrolides, some of which were conducted decades ago using bioassays,</p><p>demonstrate the largest range of penetration of all antibiotic groups (Figure 6.1). In 2</p><p>more recent studies  [28,29], patients received 500 mg azithromycin once daily for 3</p><p>days before periodontal surgery. In both studies, the average concentration ratios</p><p>increased slightly from 12 h to 2.5 days, when they reached greater than 6, and then</p><p>slowly decreased to approximately 2.5 at 6.5 days. The rate of azithromycin penetra-</p><p>tion into bone is not known, as the first samples were taken at 12 h. Telithromycin</p><p>penetration into ethmoid bone was studied in 29 patients [8]. The concentration ratio</p><p>increased between 3 and 24 h, indicating relatively slow equilibration that may favor</p><p>administration at least approximately 12 h ahead of surgery. Using naive averaging,</p><p>the average bone‐to‐serum AUC ratio was 1.6, being one of the highest extents of</p><p>penetration of all studied antibiotics.</p><p>Clindamycin</p><p>Clindamycin is often referred to as possessing high bone penetration. The median</p><p>bone‐to‐serum concentration ratio of 0.35 from four clindamycin studies, however,</p><p>appears to be lower than for fluoroquinolones (median 0.50) and linezolid (median</p><p>0.48) (Figure 6.1). As most clindamycin studies were performed in the 1970s, that is</p><p>before the introduction of fluoroquinolones, linezolid, and azithromycin, the bone</p><p>penetration of clindamycin was higher than that of other available antibiotics at that</p><p>time. All available clindamycin studies used bioassays, which have the potential to be</p><p>confounded by active metabolites of clindamycin. Results from multiple studies sug-</p><p>gest an extent of bone penetration of clindamycin of 0.21–0.45, similar to or slightly</p><p>higher than cephalosporins [5].</p><p>Rifampicin</p><p>A wide range of bone‐to‐serum concentration ratios (0.08–0.56 at 2–14 h after the dose)</p><p>was found for rifampicin in four studies in uninfected bone from the 1970s/1980s. One of</p><p>the studies also investigated infected bone, and concentration ratios were similar to those</p><p>for uninfected bone (0.57 versus 0.46). All studies utilized bioassays and had high</p><p>interpatient variability [5].</p><p>90 Bone and Joint Infections</p><p>Tetracyclines and Tigecycline</p><p>Few studies are available for tetracyclines, and results vary despite the high binding affin-</p><p>ity of tetracyclines to calcium [5]. For tigecycline, initial analysis of bone samples from 25</p><p>uninfected surgical patients suggested relatively low concentrations [9]. Re‐analysis of the</p><p>same samples by a new LC–MS/MS assay, including a stabilizing agent, resulted in bone</p><p>concentrations that were on an average 9.5‐fold higher as compared to the previous</p><p>method  [39]. This demonstrates the importance of validating all aspects of analytical</p><p>methods. Extensive penetration of tigecycline into bone has been</p><p>reported in a more</p><p>recent study in 33 uninfected surgical patients. Using a validated LC‐MS/MS assay the</p><p>authors reported a bone‐to‐serum concentration ratio of 4.77 based on the ratio of AUCs</p><p>in the matrices [40].</p><p>Cephalosporins</p><p>Numerous studies have been performed with cephalosporins. For cefuroxime, the average</p><p>bone‐to‐serum concentration ratio was 0.32 (range 0.09–0.55, 10 min to 6.5 h post dose)</p><p>in five studies that reported concentrations in serum and uninfected bone and in which</p><p>the majority of samples were above the detection limit [5]. In a recently reported study,</p><p>bone penetration of cefuroxime was studied in uninfected surgical patients undergoing</p><p>knee replacement [15]. In 9 patients who received the drug by short‐term infusion and in</p><p>9 patients administered a continuous infusion, the respective bone‐to‐plasma concentra-</p><p>tion ratios based on unbound AUCs after population PK modeling were 1.03 for cancel-</p><p>lous bone and 0.35 for cortical bone; the corresponding ratios for continuous infusion</p><p>were 1.15 and 0.65 (Table 6.2).</p><p>Ceftriaxone and cefamandole were evaluated within the same study in hip replacement</p><p>patients  [41]. At 10–30 min after the dose, average (95% confidence interval) bone‐to‐</p><p>serum concentration ratios were 0.156 (0.123–0.190) for ceftriaxone and 0.184 (0.156–</p><p>0.212) for cefamandole. The bone‐to‐plasma concentration ratios based on total drug</p><p>were similar despite a six‐fold difference in the non‐protein‐bound fractions in plasma</p><p>(0.05 for ceftriaxone, 0.30 for cefamandole), although only unbound drug is believed to</p><p>distribute between plasma and tissues. This issue is discussed in more detail in our previ-</p><p>ous review [5]. At 8 h after the dose, ceftriaxone bone‐to‐serum concentration ratios were</p><p>0.142 (0.073–0.210), very similar to those at 10–30 min, suggesting a fast equilibration</p><p>between serum and bone [41]. However, as noted in Table 6.2, a more recent study sug-</p><p>gested a slower equilibration [12]. In 11 patients undergoing debridement for septic non‐</p><p>union of the tibia, average bone‐to‐plasma AUC ratios of ceftriaxone were 0.093 in</p><p>cortical and 0.241 in cancellous bone [42].</p><p>Average bone‐to‐serum concentration ratios for cefamandole in hip replacement</p><p>patients were 0.227–0.249 at 10–30 min after the dose, and the overall range of concentra-</p><p>tion ratios reported was 0.12–2.3 at 10 min to 4 h  [5]. For cefazolin, median bone‐to‐</p><p>serum concentration ratios in eight infected patients were 0.25 (range 0.06–0.41) during</p><p>continuous infusion, with concentrations determined by bioassay [43]. In a more recent</p><p>study involving bone microdialysis and plasma sampling over 24 h, and quantification of</p><p>cefazolin by HPLC, average bone‐to‐plasma concentration ratios based on AUCs of 0.74</p><p>± 0.36 and 0.99 ± 0.59 were observed for left and right sternum of patients undergoing</p><p>coronary artery bypass grafting (CABG) [37]. Additional results for several β‐lactams are</p><p>presented in Table 6.2.</p><p>6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 91</p><p>Overall cephalosporins achieved concentration ratios of 0.1–1. Penetration was higher</p><p>into cancellous bone than into cortical bone in all studies that analyzed both, potentially</p><p>due to the higher proportion of extracellular fluid in cancellous bone  [5]. β‐Lactams,</p><p>including cephalosporins, are assumed to distribute mainly into extracellular fluid and</p><p>were found to exhibit limited binding to the inorganic bone matrix.</p><p>Penicillins, Carbapenems, and β‐Lactamase Inhibitors</p><p>Two studies by different groups from 1994 and 2001 evaluated piperacillin/tazobactam</p><p>penetration into uninfected hip bone in 12 patients each, used the same sample prepara-</p><p>tion methods and analysis by HPLC, and found consistent results. Bone‐to‐plasma con-</p><p>centration ratios were 0.2–0.3 for piperacillin and tazobactam in cortical and cancellous</p><p>bone at 1–1.5 h after the dose [44,45]. More recently, penetration of piperacillin/tazobac-</p><p>tam into uninfected jaw (n = 7) and hip (n = 2) bone was studied [46]. Sample preparation</p><p>was similar to the previous studies and concentrations were analyzed by LC‐MS/MS. At</p><p>an average of 3 h (range 1–7 h) after the start of the infusion, bone‐to‐plasma concentra-</p><p>tion ratios were 0.15 for piperacillin and 0.13 for tazobactam. These results were slightly</p><p>lower and more variable than those from the previous studies, potentially due to different</p><p>bone types and the wider range of sampling times. A wide range of average amoxicillin</p><p>bone‐to‐serum concentration ratios was reported in studies utilizing bioassays (Table 6.2).</p><p>In a study in 20 hip replacement patients analyzed by LC‐MS/MS and population PK</p><p>analysis, the bone‐to‐serum AUC ratios were 0.20 (10th–90th percentile for between‐</p><p>patient variability 0.16–0.25) for cortical and 0.18 (0.11–0.29) for cancellous bone [32]. In</p><p>the same study, the bone‐to‐serum AUC ratios of clavulanic acid were 0.15 (0.11–0.21)</p><p>for cortical and 0.10 (0.051–0.21) for cancellous bone. A recent report has documented</p><p>bone‐to‐serum concentration ratios of flucloxacillin, based on HPLC analysis, of 0.07–</p><p>0.08 and 0.05–0.06 for hip and knee bone, respectively, in the first 1.5 h after drug admin-</p><p>istration (Table 6.2) [38]. Carbapenems have been the focus in only a small number of</p><p>studies, and interpretation has been confounded by methodological problems  [5].</p><p>However, a recent study on ertapenem involving sampling over 2–28 h, analysis by LC‐</p><p>MS and population PK modeling reported a bone‐to‐plasma concentration ratio of</p><p>0.025 in 10 patients [16].</p><p>Linezolid</p><p>Linezolid is relatively stable, as opposed to many β‐lactams, and the available studies were</p><p>performed utilizing HPLC. Average (95% confidence interval) bone‐to‐serum concentra-</p><p>tion ratios were 0.51 (0.43‐0.75) in 12 hip replacement patients at 30–50 min after the</p><p>start of the infusion [47]. Similar penetration (0.40 ± 0.24) was also found at 1.5 h in 12</p><p>elderly patients during knee replacement [48]. At the same dose as the two joint replace-</p><p>ment studies [47, 48], and 0.5–1.5 h after the dose, lower linezolid concentrations were</p><p>found in 11 patients with implant‐associated infections. The average bone‐to‐plasma con-</p><p>centration ratio was approximately 0.23 [49]. However, in a recent study involving analy-</p><p>sis of linezolid by LC‐MS/MS in 9 patients with spinal tuberculosis, the average (range)</p><p>bone‐to‐plasma concentration ratio 24 h after drug administration was 0.48 (0.30–</p><p>0.67) [50]. Two studies have analyzed linezolid in bone by microdialysis [37,51]. The ratio</p><p>of unbound AUCs (fAUC) in cancellous bone/plasma over 12 h was 1.09 ± 0.11 in 3</p><p>diabetic patients with severe foot infections [51], while over 24 h in 9 uninfected patients</p><p>92 Bone and Joint Infections</p><p>undergoing CABG the ratio for sternal bone was 0.82 ± 0.28 for the left sternum and 1.02</p><p>± 0.47 for the right sternum [37]. The finding of ratios close to unity in both studies indi-</p><p>cates similar exposure to microbiologically active linezolid in interstitial bone fluid and</p><p>plasma  [37,51]. The higher AUC ratio from microdialysis, compared to concentration</p><p>ratios based on bone homogenate, is in keeping with a low propensity of linezolid to form</p><p>chelate complexes with the inorganic bone matrix.</p><p>Daptomycin</p><p>Bone penetration of daptomycin was evaluated in diabetic foot infections  [52]. Serial</p><p>microdialysis samples at steady state were collected from 0 to 8 h after the dose in 5</p><p>patients and from 8 to 16 h in another 4 patients. The average ratio of the fAUC (0–16 h)</p><p>in interstitial fluid of metatarsal bone/plasma was 1.08. This ratio of unbound bone‐to‐</p><p>unbound plasma concentrations suggests high penetration of daptomycin into intersti-</p><p>tial fluid, that is, the most likely site of infection, and was achieved for a drug with high</p><p>plasma protein binding (~90%) and a high molecular weight. Indeed, the high plasma</p><p>protein binding is the likely reason why a study that measured the ratio of total AUCs</p><p>reported a substantially lower median concentration</p><p>ratio of 0.095 for thigh bone and</p><p>0.082 for shin bone [53]. This highlights the importance of considering protein binding</p><p>when interpreting the possible clinical consequences of bone penetration results.</p><p>Fosfomycin</p><p>Bone homogenate‐to‐serum concentration ratios of 0.13–0.45 were reported in three</p><p>studies from 1980 to 1983 utilizing bioassays (Figure  6.1). Fosfomycin binds to</p><p>hydroxyapatite in bone, suggesting that not all fosfomycin in bone homogenate is micro-</p><p>biologically active. However, a microdialysis study found an average fAUC ratio of 0.43</p><p>± 0.04 in nine osteomyelitis patients with diabetic foot infection, which is higher than or</p><p>similar to the reported bone homogenate‐to‐plasma concentration ratios [54]. Considering</p><p>the very limited or no binding of fosfomycin to plasma proteins, this would suggest that</p><p>the average concentration bound to various components of bone tissue is lower than or</p><p>similar to the interstitial fluid concentrations. However, comparison among studies is</p><p>hampered by differences in study designs and methodologies.</p><p>Glycopeptides and Lipoglycopeptides</p><p>A wide range of average concentration ratios, mostly between 0.1 and 0.6, has been</p><p>reported for glycopeptides in hip, knee, or sternal bone (Figure 6.1). A recent study used</p><p>microdialysis sampling of cancellous and cortical bone to examine bone penetration of</p><p>vancomycin in 10 patients undergoing knee replacement surgery [55]. The median (95%</p><p>confidence interval) fAUC bone/plasma ratio based on samples collected up to 8 h was</p><p>0.45 (0.29‐0.62) for cancellous bone and 0.17 (0.11‐0.24) for cortical bone. The bone pen-</p><p>etration of dalbavancin was studied in 30 patients undergoing knee or hip replacement</p><p>surgery [13]. Dalbavancin was quantified in cortical bone and plasma using LC‐MS/MS,</p><p>and the resulting data subjected to population PK modeling. The bone‐to‐plasma con-</p><p>centration ratio was 0.131. When interpreting this value, it is important to recognize that</p><p>dalbavancin is approximately 90–95% bound in plasma.</p><p>6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 93</p><p>Pharmacodynamics and Monte Carlo Simulations</p><p>PD describes the relationship between drug concentrations in plasma or at the target (i.e.,</p><p>infection) site and the time‐course of drug effect(s). For β‐lactams, the time during which</p><p>the unbound antibiotic concentration remains above the minimum inhibitory concentra-</p><p>tion (fT> MIC) of the pathogen has been shown to be predictive of the extent of antibiotic</p><p>effect. For other antibiotics, such as fluoroquinolones, the fAUC/MIC best correlates</p><p>with effect.</p><p>For bone penetration studies that used the same sampling time for all samples, it is not</p><p>possible to perform a PD analysis because comparing the bone concentration at a specific</p><p>time to the MIC provides limited information. Irrespective of the type of data analysis</p><p>used, adequate reporting of the methods and assumptions is important.</p><p>Naive pooling or averaging approaches, as described earlier, can calculate the average</p><p>AUC/MIC and time above MIC. This indicates whether an “average” patient would</p><p>attain the PK/PD target. However, these naive methods have the disadvantage that they</p><p>do not consider the true variability between patients, which tends to be large for bone</p><p>penetration.</p><p>Population modeling accounts for both the average penetration and its variability</p><p>between subjects (Figure 6.2). Once a population PK model for plasma and bone has</p><p>been developed, it can be employed in Monte Carlo simulations to predict the expected</p><p>concentration‐time profiles for other than the studied dosage regimens. This includes</p><p>Serum</p><p>6</p><p>5</p><p>4</p><p>3</p><p>2</p><p>1</p><p>0</p><p>7</p><p>P25% percentile</p><p>P5% percentile</p><p>P50% percentile</p><p>P75% percentile</p><p>P95% percentile</p><p>Raw data</p><p>6</p><p>5</p><p>4</p><p>3</p><p>2</p><p>1</p><p>0</p><p>0 1 2 3 4 5 6 7</p><p>C</p><p>on</p><p>ce</p><p>nt</p><p>ra</p><p>tio</p><p>n</p><p>(m</p><p>g/</p><p>L)</p><p>C</p><p>on</p><p>ce</p><p>nt</p><p>ra</p><p>tio</p><p>n</p><p>(m</p><p>g/</p><p>kg</p><p>)</p><p>0 1 2 3</p><p>Time (h)</p><p>Time (h)</p><p>6</p><p>5</p><p>4</p><p>3</p><p>2</p><p>1</p><p>0</p><p>0 1 2 3 4 5 6 7</p><p>C</p><p>on</p><p>ce</p><p>nt</p><p>ra</p><p>tio</p><p>n</p><p>(m</p><p>g/</p><p>kg</p><p>)</p><p>Time (h)</p><p>4 5 6 7</p><p>Cortical bone</p><p>Cancellous bone</p><p>Figure 6.2. Visual predictive checks for the moxifloxacin population model presenting both the</p><p>median concentration–time profiles and their between‐subject variability in serum, cortical bone,</p><p>and cancellous bone. The diamonds represent the individual observed data from 24 hip replace-</p><p>ment patients. The solid lines represent the model‐predicted median profiles, and the dashed lines</p><p>are the model‐predicted 5, 25, 75, and 95% percentiles. (From Ref. [25], © American Society for</p><p>Microbiology).</p><p>94 Bone and Joint Infections</p><p>predicting the variability in concentration‐time profiles between patients. In this way, the</p><p>probability of achieving a PK/PD target can be predicted and recommendations made</p><p>on how to dose an antibiotic to maximize the probability of successful therapeutic</p><p>outcome.</p><p>The PK/PD target values for plasma and bone concentrations to successfully treat</p><p>bone infections are most often unknown, and the target values for other types of infec-</p><p>tions can likely not be used. For moxifloxacin, no published clinical studies in osteomy-</p><p>elitis were available. To address the lack of known PK/PD target values for bone, a reverse</p><p>engineering approach was applied for moxifloxacin to identify the most likely PK/PD</p><p>target required to achieve clinical and microbiological cure of osteomyelitis  [25]. This</p><p>approach combined effectiveness data from clinical studies with ciprofloxacin in osteo-</p><p>myelitis, the expected plasma AUCs from these studies, the AUCbone‐to‐AUCplasma</p><p>ratio for ciprofloxacin [10], fraction free (unbound) in plasma, and bacterial susceptibil-</p><p>ity data from the time of the clinical studies. Reverse engineering suggested a fAUC/MIC</p><p>of 40 in serum and an AUC/MIC of 33 in bone as the most likely PK/PD targets for</p><p>successful clinical and microbiological outcome. No assumptions were made regarding</p><p>the numerical value of the free fraction of moxifloxacin in bone. It was assumed that</p><p>binding and distribution within the bone tissue is similar for moxifloxacin and cipro-</p><p>floxacin, two quinolones with the same essential chemical structure that are expected to</p><p>be responsible for binding characteristics. The population PK model for moxifloxacin in</p><p>serum and bone was utilized to predict likely probabilities of target attainment. An ≥90%</p><p>probability of successful clinical and microbiological outcome was predicted for 400 mg</p><p>moxifloxacin once daily up to an MIC of 0.375 mg/L (mg/kg) in serum and cancellous</p><p>bone and 0.5 mg/L in cortical bone (Figure 6.3). Compared to, for example, an MIC90 of</p><p>0.125 mg/L for S. aureus, these are favorable results and suggest clinical trials are war-</p><p>ranted. The antibiotic susceptibility of the local hospital should be considered when pub-</p><p>lished probabilities of target attainment are used to decide about antibiotic therapy in</p><p>patients.</p><p>Similar analyses to determine the probability of target attainment using clinically</p><p>safe dosing regimens have been conducted for amoxicillin/clavulanic acid [32] and cip-</p><p>rofloxacin [7]. In the case of ciprofloxacin, the analysis utilized the first reported phys-</p><p>iologically based population PK model to employ measured concentrations of the</p><p>drug in plasma and cortical and cancellous bone. The population PK and Monte</p><p>Carlo simulation approach described for bone is applicable to other matrices, for</p><p>example synovial fluid.</p><p>For antibiotics for which clinical effectiveness trials are not available, combining popu-</p><p>lation PK, the reverse engineering approach utilizing effectiveness data from literature as</p><p>described earlier, and Monte Carlo simulations appears to be the best available approach</p><p>currently to derive PK/PD targets for successful treatment of bone infections and suggest</p><p>dosage regimens to be studied in clinical effectiveness trials. While sufficient bone pene-</p><p>tration is an important factor, bone concentrations alone provide limited information to</p><p>draw conclusions on the effectiveness of an antibiotic. Therefore, clinical recommenda-</p><p>tions should not be made exclusively based on bone penetration studies. An antibiotic</p><p>also needs to have adequate antibacterial activity against the infecting pathogen. Well‐</p><p>controlled PK/PD studies in osteomyelitis patients would be required to further quanti-</p><p>tatively elucidate the PK/PD relationship between antibiotic bone concentrations and</p><p>clinical outcomes. However, such studies are currently scarce.</p><p>6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 95</p><p>Conclusions</p><p>Trends in the extent of bone penetration among different groups of antibiotics have been</p><p>found from a review of greater than 140 literature studies, such as typical average pene-</p><p>tration of 0.3–1.2 for fluoroquinolones, 0.2–0.5 for linezolid, 0.1–0.3 for penicillins, and</p><p>0.1–0.5 for cephalosporins. These differences are most likely due to different physico-</p><p>chemical characteristics of the antibiotic groups. For several antibiotics, measured con-</p><p>centrations are slightly higher in cancellous bone than those in cortical bone. Developing</p><p>approaches that provide greater insights into distribution and binding of antibiotics</p><p>within bone is expected to be clinically useful. High variability between studies for a par-</p><p>ticular antibiotic group is likely partly due to a lack of standardization of bioanalytical</p><p>methods and study designs. The variability between patients within a study needs to be</p><p>considered, and this can be achieved by population PK modeling. More data are needed</p><p>to characterize the effect of infected versus uninfected bone, the presence of an implant,</p><p>and the type of bone (e.g., hip, knee, sternum) on the PK within bone. Future clinical</p><p>studies should focus on validated bioanalytical methods, as well as study designs, and</p><p>apply PK/PD analyses that incorporate the time‐course of bone concentrations to con-</p><p>tribute to evidence‐based care for patients with bone infections.</p><p>100%</p><p>P</p><p>ro</p><p>ba</p><p>bi</p><p>lit</p><p>y</p><p>of</p><p>ta</p><p>rg</p><p>et</p><p>a</p><p>tta</p><p>in</p><p>m</p><p>en</p><p>t</p><p>80%</p><p>60%</p><p>40%</p><p>20%</p><p>0%</p><p>0.125 0.25 0.5 1 2</p><p>MIC (mg/L)</p><p>Serum</p><p>4 8</p><p>100%</p><p>P</p><p>ro</p><p>ba</p><p>bi</p><p>lit</p><p>y</p><p>of</p><p>ta</p><p>rg</p><p>et</p><p>a</p><p>tta</p><p>in</p><p>m</p><p>en</p><p>t</p><p>80%</p><p>60%</p><p>40%</p><p>20%</p><p>0%</p><p>0.125 0.25 0.5 1 2</p><p>MIC (μg/g)</p><p>Cortical bone Cancellous bone</p><p>PK/PD targets for clinical and</p><p>microbiological cure based on reverse</p><p>60% engineering [25]</p><p>Serum: fAUC/MIC ≥ 40</p><p>Bone: AUC/MIC ≥ 33</p><p>4 8</p><p>100%</p><p>P</p><p>ro</p><p>ba</p><p>bi</p><p>lit</p><p>y</p><p>of</p><p>ta</p><p>rg</p><p>et</p><p>a</p><p>tta</p><p>in</p><p>m</p><p>en</p><p>t</p><p>80%</p><p>60%</p><p>40%</p><p>20%</p><p>0%</p><p>0.125 0.25 0.5 1 2</p><p>MIC (μg/g)</p><p>4 8</p><p>Figure 6.3. Probabilities of target attainment to achieve successful clinical and microbiological</p><p>outcome. The PK/PD targets for serum, cortical bone, and cancellous bone are based on a reverse</p><p>engineering approach [25].</p><p>96 Bone and Joint Infections</p><p>Key Points</p><p>● There are differences in the extent of bone penetration among various antibiotic</p><p>groups. These trends are likely related to different physicochemical and pharmacoki-</p><p>netic characteristics.</p><p>● The variability within antibiotic groups and between different studies for the same</p><p>agent is high. Therefore, utilizing standardized, validated methods for sample prepa-</p><p>ration and analysis, as well as calculating the bone‐to‐serum AUC ratios instead of</p><p>single time‐point concentration ratios, are essential.</p><p>● Based on the currently available data, combining population PK modeling, effective-</p><p>ness data from the literature, and Monte Carlo simulations appears to be the most</p><p>promising approach to elucidate the extent and time‐course of bone penetration and</p><p>its relationship with likely clinical outcomes. Well‐controlled PK/PD studies in osteomy-</p><p>elitis patients are required to directly identify PK/PD targets.</p><p>References</p><p>1. Kurtz SM, Lau EC, Son MS, et al. Are we winning or losing the battle with periprosthetic joint</p><p>infection: Trends in periprosthetic joint infection and mortality risk for the Medicare popula-</p><p>tion. J Arthroplasty. 2018;33(10):3238–3245.</p><p>2. Sendi P, Zimmerli W. Antimicrobial treatment concepts for orthopaedic device‐related infec-</p><p>tion. Clin Microbiol Infect. 2012;18(12):1176–1184.</p><p>3. Tozer TN, Rowland M. Essentials of Pharmacokinetics and Pharmacodynamics. 2nd ed.,</p><p>Philadelphia, Wolters Kluwer, 2015.</p><p>4. Drusano GL. Pharmacokinetics and pharmacodynamics of antimicrobials. Clin Infect Dis.</p><p>2007;45 Suppl 1:S89–95.</p><p>5. Landersdorfer CB, Bulitta JB, Kinzig M, et al. Penetration of antibacterials into bone: phar-</p><p>macokinetic, pharmacodynamic and bioanalytical considerations. Clin Pharmacokinet.</p><p>2009;48(2):89–124.</p><p>6. Jevon M, Guo C, Ma B, et al. Mechanisms of internalization of Staphylococcus aureus by</p><p>cultured human osteoblasts. Infect Immun. 1999;67(5):2677–2681.</p><p>7. Landersdorfer CB, Kinzig M, Höhl R, et al. Physiologically based population pharmacoki-</p><p>netic modeling approach for ciprofloxacin in bone of patients undergoing orthopedic surgery.</p><p>ACS Pharmacol Transl Sci. 2020;3(3):444–454.</p><p>8. Kuehnel TS, Schurr C, Lotter K, et al. Penetration of telithromycin into the nasal mucosa and</p><p>ethmoid bone of patients undergoing rhinosurgery for chronic sinusitis. J Antimicrob</p><p>Chemother. 2005;55(4):591–594.</p><p>9. Rodvold KA, Gotfried MH, Cwik M, et al. Serum, tissue and body fluid concentrations of</p><p>tigecycline after a single 100 mg dose. J Antimicrob Chemother. 2006;58(6):1221–1229.</p><p>10. Massias L, Buffe P, Cohen B, et al. Study of the distribution of oral ciprofloxacin into the</p><p>mucosa of the middle ear and the cortical bone of the mastoid process. Chemotherapy. 1994;40</p><p>Suppl 1:3–7.</p><p>11. Weismeier K, Adam D, Heilmann HD, et  al. Penetration of amoxycillin/clavulanate into</p><p>human bone. J Antimicrob Chemother. 1989;24 Suppl B:93–100.</p><p>12. Gergs U, Clauss T, Ihlefeld D, et al. Pharmacokinetics of ceftriaxone in plasma and bone of</p><p>patients undergoing hip or knee surgery. J Pharm Pharmacol. 2014;66(11):1552–1558.</p><p>13. Dunne MW, Puttagunta S, Sprenger CR, et al. Extended‐duration dosing and distribution of</p><p>dalbavancin into bone and articular tissue. Antimicrob Agents Chemother. 2015;59(4):</p><p>1849–1855.</p><p>6 Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone 97</p><p>14. Gergs U, Ihlefeld D, Clauss T, et al. Population pharmacokinetics of levofloxacin in plasma</p><p>and bone of patients undergoing hip or knee surgery. Clin Pharmacol Drug Dev. 2018;7(7):</p><p>692–698.</p><p>15. Tøttrup M, Søballe K, Bibby BM, et al. Bone, subcutaneous tissue and plasma pharmacokinet-</p><p>ics of cefuroxime in total knee replacement patients – a randomized controlled trial comparing</p><p>continuous and short‐term infusion. APMIS. 2019;127(12):779–788.</p><p>16. Chambers J, Page‐Sharp M, Salman S, et al. Ertapenem for osteoarticular infections in obese</p><p>patients: a pharmacokinetic study of plasma and bone concentrations. Eur J Clin Pharmacol.</p><p>2019;75(4):511–517.</p><p>17. Fong IW, Ledbetter WH, Vandenbroucke AC, et al. Ciprofloxacin concentrations in bone and</p><p>muscle after oral dosing. Antimicrob Agents Chemother. 1986;29(3):405–408.</p><p>18. Leone M, Sampol‐Manos E, Santelli D, et al. Brain tissue penetration of ciprofloxacin follow-</p><p>ing a single intravenous dose. J Antimicrob Chemother. 2002;50(4):607–609.</p><p>19. von Baum H, Böttcher S, Abel R, et al. Tissue and serum concentrations of levofloxacin in</p><p>orthopaedic patients. Int J Antimicrob Agents. 2001;18(4):335–340.</p><p>20. Rimmelé T, Boselli E, Breilh D, et al. Diffusion of levofloxacin into bone and synovial tissues.</p><p>J Antimicrob Chemother. 2004;53(3):533–535.</p><p>21. Metallidis S, Topsis D, Nikolaidis J, et al. Penetration of moxifloxacin and levofloxacin into</p><p>cancellous and cortical bone in patients undergoing total hip arthroplasty. J Chemother.</p><p>2007;19(6):682–687.</p><p>22. Meissner A, Borner K, Koeppe P. Concentrations of ofloxacin in human bone and in cartilage.</p><p>J Antimicrob Chemother. 1990;26 Suppl D:69–74.</p><p>23. Tolsdorff P. Penetration of ofloxacin into nasal tissues. Infection. 1993;21(1):66–70.</p><p>24. Tolsdorff P. Tissue and serum concentrations of ofloxacin in the ear region following a single</p><p>daily oral dose of 400 mg. Infection. 1993;21(1):63–65.</p><p>25. Landersdorfer CB, Kinzig M, Hennig FF, et al. Penetration of moxifloxacin</p><p>for the high susceptibility of implants to pyogenic infections. In subse-</p><p>quent studies in Basel, he presented experimental evidence that implants are susceptible</p><p>not only for exogenous but also for hematogenous infections. Together with Andreas F.</p><p>Widmer, he could show the special role of rifampin for the treatment of implant‐associ-</p><p>ated infections in vitro and in vivo. In conjunction with my group in Liestal, we started</p><p>to treat patients with orthopedic implant‐associated staphylococcal infections with</p><p>rifampin combinations. The promising treatment results in observational studies led to</p><p>the planning of a randomized controlled trial on the role of rifampin in patients with</p><p>implant retention. In this trial, rifampin showed its superiority in patients suffering from</p><p>acute orthopedic implant‐associated infections treated with debridement and implant</p><p>retention. After the planned interim analysis, the trial was stopped early, because all</p><p>treatment failures were observed in the arm without rifampin. This study offers one of</p><p>the few evidence‐based treatment standards in orthopedic infections.</p><p>Foreword to the First Edition</p><p>xviii Foreword to the First Edition</p><p>As an orthopedic surgeon specializing in the field of infections of the locomotor sys-</p><p>tem, I met Werner Zimmerli, a partner who is a dedicated clinician. Together with my</p><p>team in Liestal, I was in permanent contact with him to discuss our cases with bone and</p><p>joint infections. This contact was even closer when he accepted the position of Head of</p><p>the Basel University Medical Clinic in Liestal. This allowed us to create an</p><p>“Interdisciplinary Unit for Orthopedic Infections,” the first in Switzerland. Out of this</p><p>close collaboration resulted a now‐internationally respected algorithm for the treatment</p><p>of periprosthetic joint infections. Over the last two decades, a large number of Infectious</p><p>Disease specialists and orthopedic surgeons trained in the field of bone and joint infec-</p><p>tions in our group. The interdisciplinary core team later included plastic surgeons, micro-</p><p>biologists, and pathologists, a concept of collaboration that is now widely accepted.</p><p>This book reflects the concepts of interdisciplinarity. The introductory chapters deal</p><p>with general fields, which are important for managing bone and joint infections, namely</p><p>“Microbiology,” “Pharmacokinetics and Pharmacodynamics of Antibiotics in Bone,”</p><p>and “Experimental Preclinical Models.” In addition, the book contains general chapters</p><p>on “Periprosthetic Joint Infections” and “Classification of Osteomyelitis.” These chap-</p><p>ters offer to the reader detailed knowledge as a basis for competent clinical management</p><p>of bone and joint infections. The main part of the book deals with the typical clinical</p><p>entities of infections of the locomotor system. They are written following a common</p><p>concept and contain all the knowledge needed to better understand the subject treated.</p><p>Each chapter can be studied independently. Most of the chapters end with an enumera-</p><p>tion of key points and some instructive cases illustrating typical situations. This allows</p><p>the reader to assess whether he or she understood the essentials of management. Several</p><p>case examples also illustrate common errors that should be avoided. Extensive and</p><p>updated lists of references help the reader continue his or her studies. The chapters are</p><p>written by specialists with extensive clinical experience in the field of the infections that</p><p>have been described. If the treatment is mainly conservative, an infectious disease special-</p><p>ist is the author. In the chapters on infections associated with prosthetic joints and</p><p>internal fixation devices, an orthopedic surgeon joined the infectious disease specialist in</p><p>the writing team.</p><p>This book offers clear information on most problems in the management of infections</p><p>of the locomotor system. Experienced clinicians in the fields of infectious diseases, ortho-</p><p>pedic surgery, trauma surgery, rheumatology, and internal medicine can use this book as</p><p>a comprehensive textbook or on a chapter‐by‐chapter basis. Specialists in the field can</p><p>benefit from the detailed updated reviews and may find specific help for their own chal-</p><p>lenging cases.</p><p>Peter E. Ochsner</p><p>Professor Emeritus in Orthopedic</p><p>Surgery of the University of Basel</p><p>xix</p><p>Acknowledgments</p><p>I am grateful to Dr. Julia Squarr (Senior Commissioning Editor) for her support in the</p><p>planning of the new edition, as well as to Rosie Hayden, Managing Editor from John</p><p>Wiley & Sons, for her continuous help during the realization of this project. My thanks</p><p>go also to Emma Cole for her careful and efficient copyediting. This textbook would not</p><p>have been possible without the enthusiastic and competent work of the authors, who are</p><p>all specialists in different fields of bone and joint infection. My special thanks go to Ruth</p><p>Mester, who was indispensable during the whole editorial process. I would also like to</p><p>thank all the anonymous patients whose bone and joint infections were the basis of new</p><p>diagnostic and therapeutic concepts.</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>1</p><p>Chapter 1</p><p>The prevalence of most bone and joint infections is steadily increasing, mainly due to the</p><p>rising life expectancy of the population, and the increased use of bone fixation devices and</p><p>prosthetic joints. For frequent infectious diseases, such as respiratory tract, urinary tract,</p><p>and bloodstream infections, many diagnostic and therapeutic aspects have been studied in</p><p>a controlled fashion [e.g. 1–3]. In contrast, in the field of bone and joint infections, rand-</p><p>omized controlled trials are rare. Exceptions are a randomized controlled study on the role</p><p>of rifampin in patients with orthopedic implant‐associated infections, and a controlled</p><p>trial comparing two different durations of antibiotic treatment in patients with vertebral</p><p>osteomyelitis [4,5]. Therefore, diagnostic and therapeutic advice must be based mainly on</p><p>individual clinical expert knowledge and observational studies [6–10].</p><p>The optimal diagnostic and therapeutic management of bone and joint infections</p><p>needs special know‐how in different fields of medicine. Many physicians have only lim-</p><p>ited clinical experience, since arthritis and osteomyelitis are rare infectious diseases.</p><p>Therefore, a multidisciplinary approach to these infections is desirable. Only for a few</p><p>topics are internationally accepted guidelines for the management of bone and joint</p><p>infections available [11–13]. In addition, publications on the clinical practice comprising</p><p>different aspects of these infections are scarce. The aim of this book is to close this gap</p><p>with texts from a multidisciplinary team of experts in the field. Indeed, specialists in</p><p>microbiology, clinical pharmacology, preclinical research, pediatrics, pediatric and adult</p><p>orthopedic surgery, infectious diseases, and cardiovascular surgery contributed to this</p><p>book. This broad spectrum of expertise made it possible to cover a wide range of patho-</p><p>physiological, epidemiological, diagnostic, and therapeutic aspects of bone and joint</p><p>infection. The principal focus of the book is on clinical practice. It should enable clini-</p><p>cians to manage patients according to the best available evidence.</p><p>Besides the routine microbiological tests, novel non‐culture techniques are increasingly</p><p>used for the diagnosis of infectious diseases, including bone and joint infection. However,</p><p>the clinical role of molecular diagnostic procedures and mass spectrometry is ill defined.</p><p>Introduction</p><p>Werner Zimmerli</p><p>2 Bone and Joint Infections</p><p>The potential advantages of these techniques are a more rapid identification and a higher</p><p>sensitivity, especially in patients with antibiotic pretreatment or with difficult‐to‐detect</p><p>microorganisms [14,15].</p><p>With a worldwide increase in multidrug resistance, alternative antimicrobial</p><p>into bone evalu-</p><p>ated by Monte Carlo simulation. Antimicrob Agents Chemother. 2009;53(5):2074–2081.</p><p>26. Malincarne L, Ghebregzabher M, Moretti MV, et al. Penetration of moxifloxacin into bone</p><p>in patients undergoing total knee arthroplasty. J Antimicrob Chemother. 2006;57(5):</p><p>950–954.</p><p>27. Metallidis S, Charokopos N, Nikolaidis J, et al. Penetration of moxifloxacin into sternal bone</p><p>of patients undergoing routine cardiopulmonary bypass surgery. Int J Antimicrob Agents.</p><p>2006;28(5):428–432.</p><p>28. Malizia T, Tejada MR, Ghelardi E, et  al. Periodontal tissue disposition of azithromycin. J</p><p>Periodontol. 1997;68(12):1206–1209.</p><p>29. Malizia T, Batoni G, Ghelardi E, et al. Interaction between piroxicam and azithromycin during</p><p>distribution to human periodontal tissues. J Periodontol. 2001;72(9):1151–1156.</p><p>30. Grimer RJ, Karpinski MR, Andrews JM, et al. Penetration of amoxycillin and clavulanic acid</p><p>into bone. Chemotherapy. 1986;32(3):185–191.</p><p>31. Akimoto Y, Kaneko K, Tamura T. Amoxicillin concentrations in serum, jaw cyst, and jawbone</p><p>following a single oral administration. J Oral Maxillofac Surg. 1982;40(5):287–293.</p><p>32. Landersdorfer CB, Kinzig M, Bulitta JB, et al. Bone penetration of amoxicillin and clavulanic</p><p>acid evaluated by population pharmacokinetics and Monte Carlo simulation. Antimicrob</p><p>Agents Chemother. 2009;53(6):2569–2578.</p><p>33. Breilh D, Boselli E, Bel JC, et al. Diffusion of cefepime into cancellous and cortical bone tissue.</p><p>J Chemother. 2003;15(2):134–138.</p><p>34. Adam D, Reichart B, Williams KJ. Penetration of ceftazidime into human tissue in patients</p><p>undergoing cardiac surgery. J Antimicrob Chemother. 1983;12 Suppl A:269–273.</p><p>35. Raymakers JT, Schaper NC, van der Heyden JJ, et al. Penetration of ceftazidime into bone</p><p>from severely ischaemic limbs. J Antimicrob Chemother. 1998;42(4):543–545.</p><p>36. Raymakers JT, Houben AJ, van der Heyden JJ, et al. The effect of diabetes and severe ischae-</p><p>mia on the penetration of ceftazidime into tissues of the limb. Diabet Med. 2001;18(3):</p><p>229–234.</p><p>98 Bone and Joint Infections</p><p>37. Andreas M, Zeitlinger M, Wisser W, et  al. Cefazolin and linezolid penetration into sternal</p><p>cancellous bone during coronary artery bypass grafting. Eur J Cardiothorac Surg. 2015;48(5):</p><p>758–764.</p><p>38. Torkington MS, Davison MJ, Wheelwright EF, et al. Bone penetration of intravenous flucloxa-</p><p>cillin and gentamicin as antibiotic prophylaxis during total hip and knee arthroplasty. Bone</p><p>Joint J. 2017;99‐b(3):358–364.</p><p>39. Ji AJ, Saunders JP, Amorusi P, et al. A sensitive human bone assay for quantitation of tigecy-</p><p>cline using LC/MS/MS. J Pharm Biomed Anal. 2008;48(3):866–875.</p><p>40. Bhattacharya I, Gotfried MH, Ji AJ, et al. Reassessment of tigecycline bone concentrations in</p><p>volunteers undergoing elective orthopedic procedures. J Clin Pharmacol. 2014;54(1):70–74.</p><p>41. Lovering AM, Walsh TR, Bannister GC, et al. The penetration of ceftriaxone and cefamandole</p><p>into bone, fat and haematoma and relevance of serum protein binding to their penetration into</p><p>bone. J Antimicrob Chemother. 2001;47(4):483–486.</p><p>42. Garazzino S, Aprato A, Baietto L, et al. Ceftriaxone bone penetration in patients with septic</p><p>non‐union of the tibia. Int J Infect Dis. 2011;15(6):e415–421.</p><p>43. Zeller V, Durand F, Kitzis MD, et al. Continuous cefazolin infusion to treat bone and joint</p><p>infections: clinical efficacy, feasibility, safety, and serum and bone concentrations. Antimicrob</p><p>Agents Chemother. 2009;53(3):883–887.</p><p>44. Incavo SJ, Ronchetti PJ, Choi JH, et al. Penetration of piperacillin‐tazobactam into cancellous</p><p>and cortical bone tissues. Antimicrob Agents Chemother. 1994;38(4):905–907.</p><p>45. Boselli E, Breilh D, Biot L, et al. Penetration of piperacillin/tazobactam into cancellous and</p><p>cortical bone tissue. Curr Ther Res Clin Exp. 2001;62(7):538–545.</p><p>46. Al‐Nawas B, Kinzig‐Schippers M, Soergel F, et al. Concentrations of piperacillin‐tazobactam</p><p>in human jaw and hip bone. J Craniomaxillofac Surg. 2008;36(8):468–472.</p><p>47. Lovering AM, Zhang J, Bannister GC, et al. Penetration of linezolid into bone, fat, muscle and</p><p>haematoma of patients undergoing routine hip replacement. J Antimicrob Chemother.</p><p>2002;50(1):73–77.</p><p>48. Rana B, Butcher I, Grigoris P, et  al. Linezolid penetration into osteo‐articular tissues. J</p><p>Antimicrob Chemother. 2002;50(5):747–750.</p><p>49. Kutscha‐Lissberg F, Hebler U, Muhr G, et al. Linezolid penetration into bone and joint tissues</p><p>infected with methicillin‐resistant staphylococci. Antimicrob Agents Chemother. 2003;47(12):</p><p>3964–3966.</p><p>50. Li Y, Huang H, Dong W, et al. Penetration of linezolid into bone tissue 24 h after administra-</p><p>tion in patients with multidrug‐resistant spinal tuberculosis. PLoS One. 2019;14(10):e0223391.</p><p>51. Traunmüller F, Schintler MV, Spendel S, et al. Linezolid concentrations in infected soft tissue</p><p>and bone following repetitive doses in diabetic patients with bacterial foot infections. Int J</p><p>Antimicrob Agents. 2010;36(1):84–86.</p><p>52. Traunmüller F, Schintler MV, Metzler J, et  al. Soft tissue and bone penetration abilities of</p><p>daptomycin in diabetic patients with bacterial foot infections. J Antimicrob Chemother.</p><p>2010;65(6):1252–1257.</p><p>53. Montange D, Berthier F, Leclerc G, et al. Penetration of daptomycin into bone and synovial</p><p>fluid in joint replacement. Antimicrob Agents Chemother. 2014;58(7):3991–3996.</p><p>54. Schintler MV, Traunmüller F, Metzler J, et al. High fosfomycin concentrations in bone and</p><p>peripheral soft tissue in diabetic patients presenting with bacterial foot infection. J Antimicrob</p><p>Chemother. 2009;64(3):574–578.</p><p>55. Bue M, Tøttrup M, Hanberg P, et al. Bone and subcutaneous adipose tissue pharmacokinetics</p><p>of vancomycin in total knee replacement patients. Acta Orthop. 2018;89(1):95–100.</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>99</p><p>Chapter 7</p><p>Introduction</p><p>Increasing placement of medical devices, coupled with rising antibiotic resistance</p><p>amongst bacteria within the hospital and community environments, will ensure that bone</p><p>and joint infection continues to pose a major challenge to clinicians across numerous</p><p>medical specialties in the decades to come. Current treatment algorithms have benefited</p><p>from extensive preclinical studies providing in vivo evidence with regard to antimicrobial</p><p>selection and dosing. In the near future, further preclinical studies are expected to con‑</p><p>tribute vital efficacy data for new technologies such as antimicrobial loaded coatings and</p><p>vaccines, as well as rapid, sensitive, and specific diagnostics. A robust preclinical assess‑</p><p>ment of any antimicrobial strategy, and safe and expedient implementation of any such</p><p>technology, relies on a well‐designed and clinically relevant in vivo simulation using ani‑</p><p>mal models.</p><p>Many animal models of bone and joint infection have been described in the literature.</p><p>However, fully standardized, or universally accepted reference models are still lacking.</p><p>The design variables involved in creating an animal model for bone and joint infection are</p><p>multiple and inevitably require some compromise. In the field of orthopedic trauma,</p><p>where fracture‐related infection (FRI) involves additional features not commonly associ‑</p><p>ated with other types of bone and joint infection (e.g. soft tissue and vascular damage,</p><p>traumatic fracture), standardization and refinement of fixation methods is particularly</p><p>important to improve existing models and increase their clinical translational potential.</p><p>Biomechanically stable and repeatable fixation systems, which mirror clinical practice</p><p>and allow reliable healing of fractures without complication, should be the starting point</p><p>for clinically relevant research and development into anti‐infective strategies. The availa‑</p><p>bility of custom‐made fracture fixation systems for small laboratory animals [1] and the</p><p>increasing awareness of the importance of replicating the key features</p><p>of infection, such</p><p>Preclinical Models of Infection in Bone</p><p>and Joint Surgery</p><p>Caroline Constant, Lorenzo Calabro, Willem‐Jan Metsemakers, R. Geoff Richards,</p><p>and T. Fintan Moriarty</p><p>100 Bone and Joint Infections</p><p>as soft tissue involvement, in animal models are important developments in the field.</p><p>Similarly, the advent of humanized mouse models is another development with signifi‑</p><p>cant potential to improve the value of preclinical in vivo studies. These refinements and</p><p>others should enable the creation of robust, controlled, and consistent models, which</p><p>allow strong scientific conclusions with a minimum of harm to animals.</p><p>Influence of Species in Preclinical Models of Bone</p><p>and Joint Infection</p><p>Many different animal species have been used in preclinical studies as surrogates for bone</p><p>and joint infection in humans (Figure 7.1) [2]. There is currently no evidence that species</p><p>selection significantly impacts on the validity of a given study. Most studies include small</p><p>animals such as rats, rabbits, mice, and guinea pigs because of their lower cost and easier</p><p>housing and handling [3]. The central premise in performing a preclinical in vivo trial is</p><p>that the pathophysiological and therapeutic response in a chosen animal model is suffi‑</p><p>ciently similar to that in humans to allow valid extrapolation of findings.</p><p>The ideal animal model for bone and joint infection research would (i) have molecular,</p><p>cellular, structural, and mechanical features akin to human bone; (ii) have a size and</p><p>temperament allowing low‐cost maintenance and handling; (iii) have a well‐documented</p><p>genetic and immunological profile; and (iv) be sufficiently robust to endure medical and</p><p>surgical interventions that reflect current clinical practice. In reality, many of these</p><p>features are highly variable between species, and an effective study design requires aware‑</p><p>ness of these differences. The clinical reality of many bone and joint infection cases</p><p>involves chronic infection, implant loosening, long‐term antibiotic administration, and</p><p>multiple surgeries. It is clear that replicating such scenarios in preclinical animal models</p><p>would inevitably result in a high burden for the animals involved, a burden that may not</p><p>be justifiable in many preclinical studies. The extent to which a particular model is</p><p>required to reflect clinical conditions should therefore be based on a considered approach</p><p>dependent upon the research goals at hand along with consideration of the burden upon</p><p>the animal.</p><p>Historically, the success of animal models of bone and joint infection has been deter‑</p><p>mined by the degree to which radiographic, histological, and microbiological outcomes</p><p>mirror those found in human disease. Host response to infection is dependent on the</p><p>innate immunity of the host organism. The mobilization of adaptive immune defenses,</p><p>and bone modeling and remodeling mechanisms, are all species‐dependent. Bone com‑</p><p>position and micro‐ and macrostructure are important determinants of its mechanical</p><p>properties and vary between species, within species, and between anatomical loci of indi‑</p><p>vidual animals  [4]. A study by Aerssens et al.  [4] looked at the bone composition of</p><p>cows, sheep, chickens, and rats. The mineral content of femoral cortical bone was sig‑</p><p>nificantly greater in all four species than in humans, with rats having the highest con‑</p><p>tent. The proportion of collagen content showed the inverse relationship, with rats</p><p>having the lowest content. Microarchitecture of bone is related to remodeling charac‑</p><p>teristics and also differs between species. Human cortical bone, from the fetal stage</p><p>onward, shows a high degree of remodeling with a secondary osteonal structure con‑</p><p>taining concentric layers of compact bone tissue that surround a central haversian</p><p>canal. Nonhuman primates and dogs share a similar microstructure  [5], while sheep,</p><p>pigs, goats, and cows begin life with a plexiform bone structure, and develop only</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 101</p><p>secondary osteons in certain locations later in life  [6]. Rodents and rabbits have a</p><p>primary lamellar bone structure, and secondary osteons are rare [5].</p><p>Another important requirement for an animal model of bone and joint infection is the</p><p>capability of the test microorganism to cause an infection in the species under study.</p><p>Human pathogens do not necessarily cause predictable disease in a particular animal or</p><p>any disease at all. In addition, different animals may act differently to the same bacterial</p><p>inoculum [1]. Staphylococcus aureus is the pathogen cultured most frequently from clini‑</p><p>cal osteomyelitis cases [7] and is also an important cause of infection in other mammals</p><p>and domestic species [8]. Comparison of S. aureus strains from human and animal clini‑</p><p>cal isolates revealed that veterinary S. aureus infections are caused, to a large extent, by</p><p>genetically and phenotypically distinct strains [9]. The potential implication is that patho‑</p><p>genic strains have evolved separately with different immunological selection pressures</p><p>exerted by different immune defenses in different species. In addition to S. aureus, the</p><p>Tibia Radius Tibia</p><p>Figure 7.1. Examples of different animal species used in preclinical studies of bone and joint</p><p>infection at the laboratories of the co‐authors. The sheep model of chronic osteomyelitis (left) has</p><p>been used to determine the effectiveness of new local antibiotic carrier (antibiotic loaded hydrogel in</p><p>light blue) with intramedullary locking nail; the rabbit radius defect model to compare the capacity</p><p>of antibiotic‐eluting scaffolds to eradicate infection; and a rat tibial screw model to determine impact</p><p>of comorbidities on osteolysis, antibiotic efficacy and interference with repair processes.</p><p>102 Bone and Joint Infections</p><p>most common cause of osteomyelitis in domestic animals from trauma and surgical site</p><p>infection extending to the bone include S. pseudintermedius and Streptococcus spp [8].</p><p>Interspecies differences in immune response to bacterial invasion are also important to</p><p>acknowledge in infection trials. With specific reference to the causative microorganisms</p><p>involved, most animal species used in preclinical research are specific pathogen‐free, and</p><p>may have limited exposure to bacterial infection, except in case of minor cutaneous abra‑</p><p>sions allowing the skin microbiome to invade. Therefore, most animals enrolled in a bone</p><p>and joint infection study will have a very limited repertoire of circulating antibodies to</p><p>the causative pathogen. This is in contrast to the human situation, where it is considered</p><p>most people will encounter S. aureus and many coagulase‐negative staphylococci (CNS)</p><p>throughout their lifetime, and consequently have a significant adaptive immune memory</p><p>for many staphylococcal antigens. It remains to be determined what impact this has on</p><p>the progression of infection. Its relevance in the clinical situation has not been greatly</p><p>considered in preclinical research.</p><p>Murine models represent an attractive option for preclinical investigation into the host</p><p>response to infection due to the variety of genetically defined mouse strains, both wild</p><p>type and mutant, the availability of humanized strains, and the great array of molecular</p><p>different biology tools. For example, T helper (TH)–type responses are crucial in the</p><p>immune response to infection, and may be closely controlled by careful selection of</p><p>mouse strains.</p><p>Several bacterial virulence factors, including bicomponent toxins, are specific for</p><p>human receptors and represent a potential confounding factor in preclinical trials. For</p><p>example, common methicillin‐resistant S. aureus (MRSA) strains causing community‐</p><p>associated (CA)‐MRSA infection express Panton–Valentine leukocidin (PVL)  [10].</p><p>Although PVL is epidemiologically linked to severe infections in humans [11], PVL null</p><p>mutants do not reliably demonstrate reduced severity in murine models [12,13], although</p><p>it can be significantly</p><p>attenuated in rabbits [14]. Likewise, the ability of PVL to activate</p><p>or kill neutrophils is known to vary between species  [10]. The molecular basis for this</p><p>discrepancy was attributed to the high human and rabbit specificity for the PVL recep‑</p><p>tors C5aR and C5L2 [15]. The frequent recurrence of S. aureus infection and its ability to</p><p>manipulate immune responses in reaction to the pressures imposed by the human immune</p><p>system is thought to be responsible for the human‐specific activity of several potentially</p><p>important staphylococcal toxins [16]. This highlights the importance of species selection,</p><p>since some animals do not necessarily correctly replicate all facets of S. aureus disease in</p><p>humans. The recent development of immunodeficient mice reconstituted with a human</p><p>immune system [17,18], commonly called humanized mice, provide a framework to assess</p><p>the contribution of human‐specific toxins in preclinical research using a murine model of</p><p>infection [13]. These engineered mice are first made immunodeficient via gene deletion or</p><p>backcrossing strains with mutations in essential cellular partitions, such as macrophages,</p><p>natural killer, and T and B cells  [17]. Subsequently, human cells and/or tissues are</p><p>engrafted to recapitulate human immune responses  [13]. Despite the vast potential of</p><p>humanized mice, there are substantial obstacles associated with the model that remain to</p><p>be resolved to enhance transability  [17,19] and have not been substantially applied in</p><p>bone and joint infection models to date.</p><p>When evaluating bone and joint infection, the implant systems used in preclinical</p><p>research are another important factor to consider. For example, in case of a fracture, the</p><p>mechanical environment (i.e. fixation stability) is known to influence bone formation,</p><p>revascularization, and infection susceptibility [1,20]. In this regard, the use of larger animals</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 103</p><p>such as nonhuman primates, sheep, goats, and dogs can represent an advantage since</p><p>their bony geometry can accommodate human‐scale prostheses mimicking clinical</p><p>scenarios  [2]. In small animals, fixation is often performed using intramedullary</p><p>K‐wire [21–23], which may negatively impact the repair solidity compared to more stable</p><p>fixations used in clinical settings. However, small‐scale fixation devices for rodents such</p><p>as internal fixators analogous to a locking plate and interlocking nails have been devel‑</p><p>oped and have enabled small animal fracture models to better emulate clinical conditions.</p><p>These models allow the investigator to evaluate fracture healing, and choose between</p><p>rigid or flexible fixation to model primarily intramembranous or endochondral fracture</p><p>healing [24].</p><p>Overview of Animal Models</p><p>Experimental bone and joint infection models have been created in animals for diverse</p><p>purposes, by a range of means and with varied results. Common goals include profiling</p><p>infection parameters such as bacterial virulence factors or the performance of novel diag‑</p><p>nostic tools, interventions, or biomaterials. Infection is typically created by bacterial</p><p>inoculation coupled with a local perturbation in bone physiology. This can be done by an</p><p>implanted foreign body, experimentally induced ischemia, or administration of a scleros‑</p><p>ing agent. Study design in bone and joint infection research aims to reflect the clinical</p><p>situation and fit within clinical classifications (see Chapter 16).</p><p>Today, two important focus areas within the field of bone and joint infection research</p><p>are fracture‐related infection (FRI) and periprosthetic joint infection (PJI). Preclinical</p><p>models focusing on these areas should include all factors associated with orthopedic</p><p>implant‐associated infections and consider the need for revision surgery [25]. Important</p><p>FRI‐specific features to be considered for example in preclinical models include the crea‑</p><p>tion of a bone instability (fracture), soft tissue damage, and a delay in treatment (debride‑</p><p>ment and surgical fixation several hours after the accident), when mimicking an open</p><p>fracture situation. The most relevant recent preclinical in vivo models available have been</p><p>broadly categorized based upon their means of bacterial inoculation (direct/exogenous</p><p>versus hematogenous), and whether soft tissue trauma beyond the minimum required for</p><p>surgery was applied (Table 7.1) [21–23,26–38].</p><p>Direct Inoculation with Minimal Trauma</p><p>Early attempts at developing in vivo models of hematogenous osteomyelitis had found</p><p>that intravenous (IV) inoculation of bacteria alone in young healthy animals caused</p><p>inconsistent results, and that direct inoculation of S. aureus alone into intact bone failed</p><p>to create pathology mimicking chronic osteomyelitis [39]. In order to get a consistent and</p><p>progressive osteomyelitis, a sclerosing agent such as sodium morrhuate (SM) can be</p><p>administered to initiate local ischemia and tissue damage to the bone [39]. However, the</p><p>use of sclerosing agents is now considered somewhat controversial due to the unknown</p><p>effects of the sclerosing agent, which may confound results.</p><p>Currently, preclinical models show reliable results with the use of direct exogenous</p><p>bacterial inoculation within a surgical site involving bone damage and/or foreign body</p><p>placement without additional soft tissue trauma. These models are particularly suitable</p><p>Table 7.1. Select examples of animal models used for bone and joint infection research classified according to clinical situation.</p><p>Objective of research Clinical situation Species Location Type of bone instability Implant Reference</p><p>Treatment of fraction‐related</p><p>infection</p><p>Open fracture Rat Femur Bone defect Plate [26]</p><p>Close fracture Rat Femur Bone defect Plate [27]</p><p>Prevention of fraction‐related</p><p>infection</p><p>Closed fracture Rabbit Tibia Osteotomy Intramedullary [28]</p><p>Humerus Plate [29]</p><p>Rat Femur Osteotomy Plate [30]</p><p>Bone defect Plate [32]</p><p>[33]</p><p>Tibia Osteotomy Intramedullary [34]</p><p>Sheep Tibia Osteotomy Plate [35]</p><p>Open fracture Sheep Tibia Osteotomy Intramedullary [36]</p><p>Bone healing with infection Closed fracture Rat Femur Osteotomy Intramedullary [21]</p><p>Plate [31]</p><p>Osteotomies Intramedullary [23]</p><p>Pathogenesis musculoskeletal</p><p>infection</p><p>Closed fracture Mouse Femur Osteotomy Plate [37]</p><p>Rat Fracture (trauma) Intramedullary [22]</p><p>Bone defect Plate [38]</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 105</p><p>for research questions regarding PJI. Smeltzer et al. [40] showed that chronic osteomyelitis</p><p>could be generated in an otherwise healthy rabbit, by using a devascularized segment of</p><p>rabbit radius. The devascularized bone segment was inoculated with S. aureus, and it was</p><p>found that the bone served as a focus for infection propagation, which did not occur to</p><p>the same extent when the inoculum was added to an empty defect [40]. Equivalent early</p><p>work by Andriole et al.  [41] had demonstrated that a foreign body implanted in bone</p><p>played a similar role, whereby in the presence of a piece of steel within the rabbit tibiae,</p><p>an infection could develop with greater frequency than in the absence of the foreign body.</p><p>Petty et al.  [42] confirmed the high susceptibility of implants to infection. They com‑</p><p>pared different materials, and found that bone cement increased the risk for infection</p><p>more than titanium or stainless steel. However, this was probably due to heat production</p><p>during polymerization of polymethylmethacrylate. Recent studies showed that FRI models</p><p>including bone instability (osteotomy or bone defect) followed by fixation had a reliable</p><p>likelihood of infection [21,23,29–31,34,37,43–45].</p><p>Rittman and Perren [46] established an early experimental large animal model in sheep</p><p>to evaluate the impact of fixation stability on healing in an infected fracture. Their prior‑</p><p>ity, reflecting clinical focus at the time, was on histological evidence of primary bone</p><p>healing, and the model demonstrated that a fracture union was possible even when</p><p>infected under</p><p>conditions of biomechanical stability. Later studies raised concerns</p><p>regarding the safety of large animal models incorporating contaminated fractures, spe‑</p><p>cifically when using intramedullary nails [36]. More recently, a sheep infection model for</p><p>long‐bone plate osteosynthesis was developed and was used to evaluate the effect of a</p><p>hydrophobic polycationic coating, an antimicrobial‐loaded polymer sleeve, and a vancomycin‐</p><p>modified plate on biofilm formation [35,47]. Overall, this model, based on a unilat‑</p><p>eral tibial mid‐diaphyseal osteotomy repaired with an LCP plate, showed a low incidence</p><p>of complications.</p><p>Small animal models have also been described. Darouiche et al. [48] created a model,</p><p>which is noteworthy as one of the first in vivo models evaluating prophylactic implant</p><p>coatings in the presence of a fracture. A saw osteotomy was made in rabbit tibiae, which</p><p>was subsequently fixed with a chlorhexidine‐ and chloroxylenol‐coated, intramedullary</p><p>Kirschner wire and inoculated with S aureus. After six weeks, the coated nail group was</p><p>significantly less likely to have microbiological evidence of implant‐related infection than</p><p>the uncoated control group. Worlock et al.  [49] were among the first to reliably create</p><p>chronic osteomyelitis in a rabbit tibia osteotomy model. They used this model to show</p><p>the impact of fixation stability on infection susceptibility. The limitation of their experi‑</p><p>mental design was the unstable construct using a Kirschner wire in the tibia without</p><p>interlocking bolts. More stable fixation options were later investigated in rabbits; and</p><p>recent studies using intramedullary nails and plate fixation to repair bone instabilities</p><p>following osteotomies of the humerus and tibia were successfully used as FRI mod‑</p><p>els [29,44,45,50]. Osteotomy models are also widely used in rodents. Plates are success‑</p><p>fully applied following femoral osteotomies in mice to characterize the impact of S. aureus</p><p>in canaliculi of cortical bone and evaluate the immune response of fracture fixation with</p><p>and without S. aureus infection [37,43]. A similar osteotomy model was used in rats to</p><p>evaluate the necessity to remove implants following FRI, different regimens of antimi‑</p><p>crobials, and use of cationic steroid antibiotic to prevent nonunion of infected</p><p>fractures [21,23,30,34].</p><p>Rat femur models with infected large cortical bone defects have also been created,</p><p>allowing the evaluation of osteo‐inductive agents and osteo‐conductive scaffold materials</p><p>106 Bone and Joint Infections</p><p>in the context of osteomyelitis as well as efficacy of various delivery methods of antimicrobial</p><p>in contaminated fractures [27,32,33,38]. Although stabilizing these defects represented a</p><p>major challenge in the past, the recent availability of small‐scale locked plates and inter‑</p><p>locking nails have made rodents one of the main animal species used in bone and joint</p><p>infection models.</p><p>The pathogen to be inoculated is another important factor to consider in preclinical</p><p>studies. S. aureus is by far the most inoculated pathogen in animal models of FRIs at</p><p>97% [3]. In clinical reality, numerous other pathogens cause FRI; however, these other</p><p>species are currently not represented in the scientific literature [3].</p><p>Systemic antibiotics are a cornerstone in both prevention and treatment of osteomyeli‑</p><p>tis. Animal models have been used predominantly to confirm their efficacy, tailor regi‑</p><p>mens, and characterize the pharmacological parameters involved. For example [27,34],</p><p>the efficacy of rifampicin, a crucial antibiotic in the medical treatment of implant‐related</p><p>bone infection, was described in a subcutaneous tissue cage model using guinea pigs [51].</p><p>Because of rapid emergence of resistance when rifampicin is used as a single agent, a</p><p>combination antibiotic regimen for MRSA was evaluated in this model. The data showed</p><p>that daptomycin and levofloxacin are particularly effective combination partners that are</p><p>able to prevent the emergence of rifampin resistance. It should be noted that subcutane‑</p><p>ous tissue cages are commonly used as a foreign body infection model. However, this</p><p>animal model does not consider the special case of bone infection. Nevertheless, it pro‑</p><p>vides preclinical data that facilitate extrapolation to implant‐related bone infections.</p><p>Schwank et al. [58] performed an interesting approach to antibiotic therapy in preclini‑</p><p>cal testing. The novel approach involved growing bacterial biofilms on small glass beads</p><p>in vitro and exposing them to antibiotic concentrations based on normal human pharma‑</p><p>cokinetics. The authors were able to identify antibiotic combinations that could be shown</p><p>to result in eradication of biofilm in vitro, and after replicating these scenarios in the</p><p>guinea pig, there was a correlation between the regimens found to work in vitro with</p><p>clinical outcome in biofilm infections in vivo.</p><p>Numerous animal studies over the last two decades have also investigated the efficacy</p><p>of local antibiotic delivery vehicles. Technology explored using in vivo models for local</p><p>delivery of antiseptic or antibiotic agents including collagen sheets [52], calcium phos‑</p><p>phate pellets  [53], polysaccharide (chitosan)  [54], cross‐linked high amylose starch</p><p>implants [55], biodegradable polymer beads [56], implant coatings [45], covalently bonded</p><p>antibiotics [57], and gel or thermo‐responsive vehicles [27,29]. In this situation, the in vivo</p><p>models used are designed to analyze outcomes such as drug release profiles, biocompat‑</p><p>ibility, effects on fracture healing, infection susceptibility, and drug resistance.</p><p>A rat model designed by Lucke et al. [58] to evaluate a gentamicin‐impregnated poly‐d,l‐</p><p>lactic acid (PDLLA)–coated nail, which has since been approved for clinical use, provides</p><p>a good example of translational research. Kirschner wires were coated with a PDLLA</p><p>polymer containing gentamicin and implanted in the tibial medullary canal along with S.</p><p>aureus. There was a significant reduction in clinical symptoms in the treatment group</p><p>compared with controls. In follow‐up experiments, using the same rat model, the antibi‑</p><p>otic burst release profile and osseous drug concentrations at progressive time points were</p><p>described. This study was conducted in parallel with a prospective clinical trial of the nail</p><p>in open tibial fractures [59]. This particular antibiotic‐containing medical device is only</p><p>one of many currently being developed, but this series of studies demonstrates the inte‑</p><p>gral role of the animal model in proving the efficacy of the coating prior to successful</p><p>introduction into the clinic.</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 107</p><p>Successfully treating or managing an established active bone or joint infection is a distinctly</p><p>more challenging undertaking than preventing infection. Biofilm formation, poor osse‑</p><p>ous perfusion of antibiotics, and abscess formation necessitate a multimodal medical and</p><p>surgical approach. In vivo models for treatment of infection are correspondingly more</p><p>complex. At least two surgical procedures are required – one to inoculate, and the second</p><p>to treat, often requiring surgical site debridement (Figure 7.2), with an intervening period</p><p>to allow infection to develop. Interventions effective in prophylaxis will not necessarily be</p><p>effective in treatment. For example, using a canine bone infection model, it was shown</p><p>that PMMA loaded with gentamicin was successful in preventing the development of an</p><p>infection. However, it was unable to successfully treat an active infection [60]. The clinical</p><p>conditions surrounding the treatment of an active infection, including biofilm formation,</p><p>intracellular bacteria, and tissue necrosis, represent a significantly more challenging</p><p>target for any antibiotic‐loaded biomaterial.</p><p>Figure 7.2. Mouse infection model with bone instability and locking plate placed during a first</p><p>surgical procedure with bacterial inoculation that underwent a second surgical procedure for</p><p>debridement</p><p>of surgical site that showed an abscess (arrow).</p><p>108 Bone and Joint Infections</p><p>Animal Models of Bone and Joint Infection Incorporating Trauma</p><p>Infection following open fractures causes significant clinical morbidity and creates unique</p><p>management challenges [25]. Soft tissue injury in trauma patients is known to increase</p><p>the FRI rate [61] and to significantly increase infection rates after a standardized closed</p><p>soft tissue injury in rats [62]. However, traumatic fractures and soft tissue injuries add an</p><p>additional level of complexity to infection models and are incorporated in only 25% of</p><p>studies, primarily due to the burden upon the experimental animal [3]. As described ear‑</p><p>lier, most researchers performed an osteotomy or bone defect to simulate a fracture.</p><p>Although creating a traumatic (“true”) fracture is more realistic, creating an osteotomy</p><p>decreases confounding factors from fracture configuration, improves reproducibility of</p><p>in vivo models, and keeps the number of animals to a minimum. Furthermore, fracture</p><p>stability can be more difficult to achieve in traumatic fractures compared to a controlled</p><p>osteotomy. Regardless, creating a real fracture is far more realistic and more comparable</p><p>to clinical situations where fractures are often accompanied by soft tissue damage, hema‑</p><p>tomas, and potential vascular compromise. Techniques used to create a fracture associ‑</p><p>ated with soft tissue trauma in preclinical models include blunt trauma from a weight</p><p>being dropped on the bone [63–66], three‐point bending apparatus to mimic traumatic</p><p>forces  [67], firing a steel fragment to the tibia  [68] or with a haemostat  [69]. Lindsey</p><p>et al. [63] reported on a rat femur model, where they were able to create a reproducible</p><p>closed fracture with a blunt impact guillotine, fix it with a Kirschner wire, and inoculate</p><p>S. aureus to create local infection without overwhelming sepsis. This model was later used</p><p>to demonstrate the importance of timing of antibiotic administration and surgery regard‑</p><p>ing the rate of infection in wounds contaminated with S. aureus. By varying the timing of</p><p>both interventions, it was shown that early antibiotic therapy was the single most impor‑</p><p>tant factor for the risk of infection [70]. A delay in surgery did result in an increase in</p><p>infection rate, though there was no significant increase between a delay of 6 and 24 h.</p><p>Only one preclinical model of FRI including all clinical features of the condition (frac‑</p><p>ture creation, soft tissue damage, delay in treatment) was described. This model uses</p><p>Sprague‐Dawley rats and creates a femoral fracture using a blunt trauma from a weight</p><p>dropped from a height. Following fracture creation, the surgical site is inoculated with a</p><p>S. aureus bacterial suspension and left open for 1h prior to surgical stabilization using an</p><p>intramedullary K‐wire [63‑66]. The potential ethical issues and complications inherent in</p><p>these models are exemplified in a recent study, where efforts were made to reliably pro‑</p><p>duce a traumatic tibial fracture in a rat model  [71]. Subject animals were divided into</p><p>small groups depending on fracture configuration and whether the fracture was fixed</p><p>with an intramedullary pin or with external coaptation. The authors reported extremity</p><p>necrosis at day 7 postoperatively and several animals showing weight loss and signs of</p><p>obvious malnutrition.</p><p>Hematogenous Models</p><p>Hematogenous osteomyelitis is a particular issue in pediatric medicine where septic</p><p>arthritis and infection in the adjacent metaphyses of long bones is relatively common [72].</p><p>It is also a common cause for late infection in previously well‐functioning prosthetic</p><p>joints, with a 39% risk of PJI following a S. aureus bacteremia [73]. Acknowledging this</p><p>distinct etiology, a number of authors have attempted to create models of hematogenous</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 109</p><p>osteomyelitis using S. aureus, and a focal bone lesion typically created either concurrently</p><p>or prior to inoculation. An early model, created by Deysine et al. [74], involved injecting</p><p>a combination of barium and 5 × 105 CFU of S. aureus simultaneously into the tibial</p><p>nutrient artery of dogs. The authors reliably created acute osteomyelitis, but there was an</p><p>unacceptably high mortality rate from sepsis. Other similar models in chickens [75] and</p><p>rabbits [76] encountered the same problem, describing narrow safety margins in inocu‑</p><p>lum dose when administered systemically. Hienz et al. [77], however, were able to create a</p><p>model using rats with no mortality during the 14‐day study period. They injected SM</p><p>locally into the mandible and tibia and inoculated varying doses of S. aureus intrave‑</p><p>nously into the tail vein to determine the dose required to infect 50% (ID50) and 100%</p><p>(ID100) of animals. Animals inoculated with bacteria, but spared the focal SM injection,</p><p>did not develop osteomyelitis, again demonstrating the importance of local perturbation</p><p>of bone physiology. An alternative approach was taken by Whalen et al. [78] who designed</p><p>a model to mimic pediatric hematogenous osteomyelitis. Using skeletally immature rab‑</p><p>bits, they created a partial growth plate fracture at the proximal tibial metaphysis and</p><p>administered S. aureus intravenously via an ear vein, reliably causing focal acute osteo‑</p><p>myelitis without metastatic infection. Skeletally immature (20–24 weeks) rabbits were</p><p>used in this study, in order to simulate pediatric hematogenous osteomyelitis. The impor‑</p><p>tance of choosing a model replicating clinical cases was highlighted by a study using a</p><p>model of localized osteomyelitis by injecting S. aureus unilaterally into the femoral artery</p><p>of domestic pigs [79]. The use of juvenile domestic pigs (40 kg) led to early euthanasia of</p><p>several animals due to lameness, shallow respiration, fever, and anorexia in contrast to</p><p>fewer complications and euthanasia in younger pigs (20 kg). In addition, osteomyelitis</p><p>caused by S. aureus was seen in six of the seven 20 kg pigs and in none of the 40 kg</p><p>pigs [80]. More recently, a murine hematogenous osteomyelitis model that closely mimics</p><p>the human infection was described [81,82]. The mice were infected intravenously with 106</p><p>CFUs of S. aureus via the tail vein and subsequently developed chronic osteomyelitis.</p><p>Future Directions</p><p>Animal models of bone and joint infection are developed with the primary aim of</p><p>improving outcomes in clinical medicine. In theory, they allow in vivo evaluation of</p><p>potential therapies, prophylactics, and diagnostics without the costs, safety, and ethical</p><p>issues associated with human clinical trials.</p><p>Prophylactic strategies in orthopedic surgery, such as systemic and local antibiotic</p><p>therapy, have evolved over the last four decades. Clinical use of IV antibiotics and antibi‑</p><p>otic‐loaded bone cement in arthroplasty grew sporadically throughout the 1970s and was</p><p>followed only later with detailed characterization in controlled animal models  [83].</p><p>Mainstay surgical techniques in the treatment of established infection such as debride‑</p><p>ment, stabilization, and lavage have also been characterized after the fact, rather than</p><p>developed with the help of animal models. The limitations involved with these traditional</p><p>strategies, and an ever‐expanding understanding of bacterial virulence factors, have,</p><p>however, fueled significant efforts to develop new technology to combat osteomyelitis,</p><p>which inevitably increases the demand for reliable and valid in vivo models [84].</p><p>Promising new approaches to prevent infection include (i) modification of implant</p><p>surface to deter bacterial adhesion [57]; (ii) coating implants with degradable polymers</p><p>that elute high concentrations of antibiotic into the local milieu (without causing toxic</p><p>110 Bone and Joint Infections</p><p>systemic concentrations) [58,85]; (iii) new drugs targeted at either the genes or effector</p><p>molecules of adhesion [86], quorum sensing [87] or RNA processing [88]; and (iv) vaccine</p><p>development against</p><p>biofilm‐forming bacteria [89].</p><p>Conclusions</p><p>Animal models in modern biomedical research are indispensable in the development of</p><p>novel interventional and diagnostic technologies. The success of any future anti‐infective</p><p>technology will depend upon proper evaluation in appropriate animal models. Robust</p><p>assessment of the performance of any clinical device may require testing in comparatively</p><p>low‐burden animal models, although higher‐burden models, including, for example, frac‑</p><p>ture creation and localized tissue damage, will be required in certain cases. The develop‑</p><p>ment of refined small animal models will enable screening of candidate technologies that</p><p>are gated at an early stage to reduce the need for more burdensome investigations of any</p><p>but the most promising candidates. Real‐time, in vivo estimation of bacterial burden is also</p><p>likely to be a key area for a reduction in the number of animals required in the future.</p><p>Preclinical models investigating musculoskeletal infection should recapitulate specific</p><p>features of the clinical condition [2,26]. Current preclinical in vivo musculoskeletal infec‑</p><p>tion models rarely mimic a clinical situation of musculoskeletal infection. In addition to</p><p>the overrepresentation of S. aureus in infection models, other clinically important factors</p><p>such as soft tissue trauma are rarely recreated. Refinement of existing preclinical models</p><p>and development of new models more resembling to clinical scenarios are needed to</p><p>make progress in the field of prevention and treatment of bone and joint infection.</p><p>Furthermore, the recent development of humanized are a promising avenue to increase</p><p>translatability of preclinical research using murine model of infection [13].</p><p>Key Points</p><p>● Different animal species may vary in bone structure, susceptibility to infection, adaptive</p><p>immune response to bacteria, and specificity of bacterial toxins.</p><p>● The implant systems available for laboratory animals are improving, with biomechan‑</p><p>ically defined fracture fixation now available in rodent models.</p><p>● Investigators should weigh the importance of clinical relevance versus burden upon</p><p>the animal when deciding upon the particular model chosen.</p><p>● Testing novel prophylactic measures requires different models in comparison with</p><p>testing of novel treatments for bone and joint infection.</p><p>● The use of immunodeficient mice reconstituted with a human hematopoietic immune</p><p>system allows assessment of human‐specific toxins in murine model of infection.</p><p>References</p><p>1. Sabaté Brescó M, O’Mahony L, Zeiter S, et al. Influence of fracture stability on Staphylococcus</p><p>epidermidis and Staphylococcus aureus infection in a murine femoral fracture model. Eur Cell</p><p>Mater. 2017;34:321–340.</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 111</p><p>2. Moriarty TF, Harris LG, Mooney RA, et al. Recommendations for design and conduct of</p><p>preclinical in vivo studies of orthopedic device‐related infection. Journal of Orthopaedic</p><p>Research. 2019;37(2):271–287.</p><p>3. Vanvelk N, Morgenstern M, Moriarty TF, et al. Preclinical in vivo models of fracture‐related</p><p>infection: a systematic review and critical appraisal. Eur Cell Mater. 2018;36:184–199.</p><p>4. Aerssens J, Boonen S, Lowet G, et al. Interspecies differences in bone composition, density, and</p><p>quality: potential implications for in vivo bone research. Endocrinology. 1998;139(2):</p><p>663–670.</p><p>5. Currey JD. Bones: structure and mechanics: Princeton University Press, 2006.</p><p>6. Reinwald S, Burr D. Review of nonprimate, large animal models for osteoporosis research.</p><p>J Bone Miner Res. 2008;23(9):1353–1368.</p><p>7. Torbert JT, Joshi M, Moraff A, et al. Current bacterial speciation and antibiotic resistance in</p><p>deep infections after operative fixation of fractures. J Orthop Trauma. 2015;29(1):7–17.</p><p>8. Gieling F, Peters S, Erichsen C, et  al. Bacterial osteomyelitis in veterinary orthopaedics:</p><p>Pathophysiology, clinical presentation and advances in treatment across multiple species. The</p><p>Veterinary Journal. 2019;250:44–54.</p><p>9. Sung JM, Lloyd DH, Lindsay JA. Staphylococcus aureus host specificity: comparative genom‑</p><p>ics of human versus animal isolates by multi‐strain microarray. Microbiology (Reading).</p><p>2008;154(Pt 7):1949–1959.</p><p>10. Löffler B, Hussain M, Grundmeier M, et al. Staphylococcus aureus panton‐valentine leukoci‑</p><p>din is a very potent cytotoxic factor for human neutrophils. PLoS Pathog. 2010;6(1):e1000715.</p><p>11. Gillet Y, Issartel B, Vanhems P, et al. Association between Staphylococcus aureus strains carry‑</p><p>ing gene for Panton‐Valentine leukocidin and highly lethal necrotising pneumonia in young</p><p>immunocompetent patients. Lancet. 2002;359(9308):753–759.</p><p>12. Shallcross LJ, Fragaszy E, Johnson AM, et al. The role of the Panton‐Valentine leucocidin</p><p>toxin in staphylococcal disease: a systematic review and meta‐analysis. Lancet Infect Dis.</p><p>2013;13(1):43–54.</p><p>13. Prince A, Wang H, Kitur K, et al. Humanized mice exhibit increased susceptibility to Staphylococcus</p><p>aureus pneumonia. The Journal of Infectious Diseases. 2016;215(9):1386–1395.</p><p>14. Lipinska U, Hermans K, Meulemans L, et al. Panton‐Valentine leukocidin does play a role in</p><p>the early stage of Staphylococcus aureus skin infections: a rabbit model. PLoS One. 2011;</p><p>6(8):e22864.</p><p>15. Spaan AN, Henry T, van Rooijen WJM, et  al. The staphylococcal toxin Panton‐Valentine</p><p>Leukocidin targets human C5a receptors. Cell Host Microbe. 2013;13(5):584–594.</p><p>16. Thammavongsa V, Kim HK, Missiakas D, et al. Staphylococcal manipulation of host immune</p><p>responses. Nat Rev Microbiol. 2015;13(9):529–543.</p><p>17. Yong KSM, Her Z, Chen Q. Humanized mice as unique tools for human‐specific studies. Arch</p><p>Immunol Ther Exp (Warsz). 2018;66(4):245–266.</p><p>18. Shultz LD, Brehm MA, Garcia‐Martinez JV, et al. Humanized mice for immune system inves‑</p><p>tigation: progress, promise and challenges. Nat Rev Immunol. 2012;12(11):786–798.</p><p>19. Shi Y, Inoue H, Wu JC, et al. Induced pluripotent stem cell technology: a decade of progress.</p><p>Nat Rev Drug Discov. 2017;16(2):115–130.</p><p>20. Claes L, Meyers N, Schülke J, et al. The mode of interfragmentary movement affects bone</p><p>formation and revascularization after callus distraction. PLOS ONE. 2018;13(8):e0202702.</p><p>21. Bilgili F, Balci HI, Karaytug K, et  al. Can normal fracture healing be achieved when the</p><p>implant is retained on the basis of infection? An experimental animal model. Clin Orthop Relat</p><p>Res. 2015;473(10):3190–3196.</p><p>22. Gilbert SR, Camara J, Camara R, et al. Contaminated open fracture and crush injury: a murine</p><p>model. Bone Res. 2015;3:14050–14050.</p><p>23. Schindeler A, Yu NY, Cheng TL, et al. Local delivery of the cationic steroid antibiotic CSA‐90</p><p>enables osseous union in a rat open fracture model of Staphylococcus aureus infection. J Bone</p><p>Joint Surg Am. 2015;97(4):302–309.</p><p>112 Bone and Joint Infections</p><p>24. Histing T, Garcia P, Matthys R, et al. An internal locking plate to study intramembranous</p><p>bone healing in a mouse femur fracture model. J Orthop Res. 2010;28(3):397–402.</p><p>25. Depypere M, Morgenstern M, Kuehl R, et  al. Pathogenesis and management of fracture‐</p><p>related infection. Clinical Microbiology and Infection. 2020;26(5):572–578.</p><p>26. Brown KV, Penn‐Barwell JG, Rand BC, et al. Translational research to improve the treatment</p><p>of severe extremity injuries. Journal of the Royal Army Medical Corps. 2014;160(2): 167–170.</p><p>27. Rand BC, Penn‐Barwell JG, Wenke JC. Combined local and systemic antibiotic delivery</p><p>improves eradication of wound contamination: An animal experimental model of contami‑</p><p>nated fracture. Bone Joint J. 2015;97‐b(10):1423–1427.</p><p>28. Li Y, Chen S‐K, Li L, et al. Bone defect animal models for testing efficacy of bone substitute</p><p>biomaterials. Journal of Orthopaedic Translation. 2015;3(3):95–104.</p><p>29. Ter Boo GA, Arens D, Metsemakers WJ, et al. Injectable gentamicin‐loaded thermo‐respon‑</p><p>sive hyaluronic acid derivative prevents infection in a rabbit model. Acta Biomater.</p><p>2016;43:185–194.</p><p>30. Lovati AB, Drago L, Bottagisio M, et al. Systemic and local</p><p>administration of antimicrobial</p><p>and cell therapies to prevent methicillin‐resistant Staphylococcus epidermidis‐induced femoral</p><p>nonunions in a rat model. Mediators Inflamm. 2016;2016:9595706.</p><p>31. Lovati AB, Romanò CL, Bottagisio M, et  al. Modeling Staphylococcus epidermidis‐</p><p>induced non‐unions: Subclinical and clinical evidence in rats. PLoS One. 2016;11(1):</p><p>e0147447.</p><p>32. Penn‐Barwell JG, Baker B, Wenke JC. Local bismuth thiols potentiate antibiotics and reduce</p><p>infection in a contaminated open fracture model. J Orthop Trauma. 2015;29(2):e73–78.</p><p>33. Tennent DJ, Shiels SM, Sanchez CJ, Jr., et al. Time‐dependent effectiveness of locally applied</p><p>vancomycin powder in a contaminated traumatic orthopaedic wound model. J Orthop Trauma.</p><p>2016;30(10):531–537.</p><p>34. Sethi S, Thormann U, Sommer U, et al. Impact of prophylactic CpG Oligodeoxynucleotide applica‑</p><p>tion on implant‐associated Staphylococcus aureus bone infection. Bone. 2015;78:194–202.</p><p>35. Schaer TP, Stewart S, Hsu BB, et al. Hydrophobic polycationic coatings that inhibit biofilms</p><p>and support bone healing during infection. Biomaterials. 2012;33(5):1245–1254.</p><p>36. Hill PF, Clasper JC, Parker SJ, et al. Early intramedullary nailing in an animal model of a heav‑</p><p>ily contaminated fracture of the tibia. J Orthop Res. 2002;20(4):648–653.</p><p>37. de Mesy Bentley KL, Trombetta R, Nishitani K, et  al. Evidence of Staphylococcus aureus</p><p>deformation, proliferation, and migration in canaliculi of live cortical bone in murine models</p><p>of osteomyelitis. J Bone Miner Res. 2017;32(5):985–990.</p><p>38. Seebach E, Holschbach J, Buchta N, et al. Mesenchymal stromal cell implantation for stimula‑</p><p>tion of long bone healing aggravates Staphylococcus aureus induced osteomyelitis. Acta</p><p>Biomater. 2015;21:165–177.</p><p>39. Schrman L, Janota M, Lewin P. The production of experimental osteomyelitis: preliminary</p><p>report. Journal of the American Medical Association. 1941;117(18):1525–1529.</p><p>40. Smeltzer MS, Thomas JR, Hickmon SG, et al. Characterization of a rabbit model of staphylo‑</p><p>coccal osteomyelitis. J Orthop Res. 1997;15(3):414–421.</p><p>41. Andriole VT, Nagel DA, Southwick WO. A paradigm for human chronic osteomyelitis. J Bone</p><p>Joint Surg Am. 1973;55(7):1511–1515.</p><p>42. Petty W, Spanier S, Shuster JJ, et al. The influence of skeletal implants on incidence of infec‑</p><p>tion. Experiments in a canine model. J Bone Joint Surg Am. 1985;67(8):1236–1244.</p><p>43. Rochford ETJ, Sabaté Brescó M, Zeiter S, et al. Monitoring immune responses in a mouse</p><p>model of fracture fixation with and without Staphylococcus aureus osteomyelitis. Bone.</p><p>2016;83:82–92.</p><p>44. Arens D, Wilke M, Calabro L, et al. A rabbit humerus model of plating and nailing osteosyn‑</p><p>thesis with and without Staphylococcus aureus osteomyelitis. Eur Cell Mater. 2015;30:148–161;</p><p>discussion 61–62.</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 113</p><p>45. Prinz C, Elhensheri M, Rychly J, et al. Antimicrobial and bone‐forming activity of a copper</p><p>coated implant in a rabbit model. J Biomater Appl. 2017;32(2):139–149.</p><p>46. Rittmann W, Perren S. Cortical bone healing after internal fixation and fracture. Springer/</p><p>Verlag. Berlin/Heidelberg/New York, 1974.</p><p>47. Stewart S, Barr S, Engiles J, et al. Vancomycin‐modified implant surface inhibits biofilm forma‑</p><p>tion and supports bone‐healing in an infected osteotomy model in sheep: a proof‐of‐concept</p><p>study. The Journal of bone and joint surgery American volume. 2012;94(15):1406–1415.</p><p>48. Darouiche RO, Farmer J, Chaput C, et al. Anti‐infective efficacy of antiseptic‐coated intramed‑</p><p>ullary nails. J Bone Joint Surg Am. 1998;80(9):1336–1340.</p><p>49. Worlock P, Slack R, Harvey L, et al. The prevention of infection in open fractures. An experi‑</p><p>mental study of the effect of antibiotic therapy. J Bone Joint Surg Am. 1988;70(9):1341–1347.</p><p>50. Metsemakers WJ, Schmid T, Zeiter S, et al. Titanium and steel fracture fixation plates with dif‑</p><p>ferent surface topographies: Influence on infection rate in a rabbit fracture model. Injury.</p><p>2016;47(3):633–639.</p><p>51. John AK, Baldoni D, Haschke M, et al. Efficacy of daptomycin in implant‐associated infection</p><p>due to methicillin‐resistant Staphylococcus aureus: importance of combination with rifampin.</p><p>Antimicrob Agents Chemother. 2009;53(7):2719–2724.</p><p>52. Riegels‐Nielsen P, Espersen F, Hölmich LR, et al. Collagen with gentamicin for prophylaxis of</p><p>postoperative infection. Staphylococcus aureus osteomyelitis studied in rabbits. Acta Orthop</p><p>Scand. 1995;66(1):69–72.</p><p>53. Beardmore AA, Brooks DE, Wenke JC, et al. Effectiveness of local antibiotic delivery with an</p><p>osteoinductive and osteoconductive bone‐graft substitute. J Bone Joint Surg Am. 2005;</p><p>87(1):107–112.</p><p>54. Paiva Costa L, Moreira Teixeira LE, Maranhão Lima GS, et  al. Effectiveness of Chitosan</p><p>Films Impregnated With Ciprofloxacin for the Prophylaxis of Osteomyelitis in Open Fractures:</p><p>An Experimental Study in Rats. Arch Trauma Res. 2016;5(3):e36952–e52.</p><p>55. Huneault LM, Lussier B, Dubreuil P, et al. Prevention and treatment of experimental osteomy‑</p><p>elitis in dogs with ciprofloxacin‐loaded crosslinked high amylose starch implants. J Orthop Res.</p><p>2004;22(6):1351–1357.</p><p>56. Ambrose CG, Clyburn TA, Louden K, et al. Effective treatment of osteomyelitis with biode‑</p><p>gradable microspheres in a rabbit model. Clin Orthop Relat Res. 2004. doi: 10.1097/01.</p><p>blo.0000126303.41711.a2(421):293‐9.</p><p>57. Antoci V, Jr., Adams CS, Parvizi J, et al. The inhibition of Staphylococcus epidermidis biofilm</p><p>formation by vancomycin‐modified titanium alloy and implications for the treatment of</p><p>periprosthetic infection. Biomaterials. 2008;29(35):4684–4690.</p><p>58. Lucke M, Schmidmaier G, Sadoni S, et al. Gentamicin coating of metallic implants reduces</p><p>implant‐related osteomyelitis in rats. Bone. 2003;32(5):521–531.</p><p>59. Fuchs T, Stange R, Schmidmaier G, et al. The use of gentamicin‐coated nails in the tibia: pre‑</p><p>liminary results of a prospective study. Arch Orthop Trauma Surg. 2011;131(10): 1419–1425.</p><p>60. Fitzgerald RH, Jr. Experimental osteomyelitis: description of a canine model and the role of</p><p>depot administration of antibiotics in the prevention and treatment of sepsis. J Bone Joint Surg</p><p>Am. 1983;65(3):371–380.</p><p>61. Young K, Aquilina A, Chesser TJS, et al. Open tibial fractures in major trauma centres: A</p><p>national prospective cohort study of current practice. Injury. 2019;50(2):497–502.</p><p>62. Kälicke T, Schlegel U, Printzen G, et al. Influence of a standardized closed soft tissue trauma</p><p>on resistance to local infection. An experimental study in rats. J Orthop Res. 2003;21(2):</p><p>373–378.</p><p>63. Lindsey BA, Clovis NB, Smith ES, et al. An animal model for open femur fracture and osteo‑</p><p>myelitis: Part I. J Orthop Res. 2010;28(1):38–42.</p><p>64. Boyce BM, Lindsey BA, Clovis NB, et al. Additive effects of exogenous IL‐12 supplementation</p><p>and antibiotic treatment in infection prophylaxis. J Orthop Res. 2012;30(2):196–202.</p><p>114 Bone and Joint Infections</p><p>65. Li B, Jiang B, Boyce BM, et al. Multilayer polypeptide nanoscale coatings incorporating IL‐12</p><p>for the prevention of biomedical device‐associated infections. Biomaterials. 2009;30(13):</p><p>2552–2558.</p><p>66. Lindsey BA, Clovis NB, Smith ES, et al. An animal model for open femur fracture and osteo‑</p><p>myelitis—Part II: Immunomodulation with systemic IL‐12. Journal of Orthopaedic Research.</p><p>2010;28(1):43–47.</p><p>67. Ashhurst DE, Hogg J, Perren SM. A method for making reproducible experimental fractures</p><p>of the rabbit tibia. Injury. 1982;14(3):236–242.</p><p>68. Hill PF, Watkins PE. The prevention of experimental osteomyelitis in a model of gunshot</p><p>fracture in the pig. European Journal of Orthopaedic Surgery & Traumatology. 2001;11(4):</p><p>237–241.</p><p>69. Petri WH, 3rd, Schaberg SJ. The effects of antibiotic‐supplemented bone allografts on con‑</p><p>taminated, partially avulsive fractures of the canine ulna. J Oral Maxillofac Surg. 1984;</p><p>42(11):699–704.</p><p>70. Penn‐Barwell JG, Murray CK, Wenke JC. Early antibiotics and debridement independently</p><p>reduce infection in an open fracture model. J Bone</p><p>Joint Surg Br. 2012;94(1):107–112.</p><p>71. Shi E, Chen G, Qin B, et al. A novel rat model of tibial fracture for trauma researches: a com‑</p><p>bination of different types of fractures and soft tissue injuries. J Orthop Surg Res. 2019;14(1):333.</p><p>72. Brown DW, Sheffer BW. Pediatric septic arthritis: An update. Orthop Clin North Am.</p><p>2019;50(4):461–470.</p><p>73. Sendi P, Banderet F, Graber P, et al. Periprosthetic joint infection following Staphylococcus</p><p>aureus bacteremia. J Infect. 2011;63(1):17–22.</p><p>74. Deysine M, Rosario E, Isenberg HD. Acute hematogenous osteomyelitis: an experimental</p><p>model. Surgery. 1976;79(1):97–99.</p><p>75. Emslie KR, Nade S. Pathogenesis and treatment of acute hematogenous osteomyelitis: evaluation</p><p>of current views with reference to an animal model. Rev Infect Dis. 1986;8(6): 841–849.</p><p>76. Poultsides LA, Papatheodorou LK, Karachalios TS, et al. Novel model for studying hematog‑</p><p>enous infection in an experimental setting of implant‐related infection by a community‐</p><p>acquired methicillin‐resistant S. aureus strain. J Orthop Res. 2008;26(10):1355–1362.</p><p>77. Hienz SA, Sakamoto H, Flock JI, et al. Development and characterization of a new model of</p><p>hematogenous osteomyelitis in the rat. J Infect Dis. 1995;171(5):1230–1236.</p><p>78. Whalen JL, Fitzgerald RH, Jr., Morrissy RT. A histological study of acute hematogenous</p><p>osteomyelitis following physeal injuries in rabbits. J Bone Joint Surg Am. 1988;70(9):</p><p>1383–1392.</p><p>79. Nielsen OL, Afzelius P, Bender D, et al. Comparison of autologous (111)In‐leukocytes, (18)</p><p>F‐FDG, (11)C‐methionine, (11)C‐PK11195 and (68)Ga‐citrate for diagnostic nuclear imaging</p><p>in a juvenile porcine haematogenous staphylococcus aureus osteomyelitis model. Am J Nucl</p><p>Med Mol Imaging. 2015;5(2):169–182.</p><p>80. Alstrup AKO, Nielsen KM, Schønheyder HC, et al. Refinement of a hematogenous localized</p><p>osteomyelitis model in pigs. Scand J Lab Anim Sci. 2016;42:1–4.</p><p>81. Horst SA, Hoerr V, Beineke A, et al. A novel mouse model of Staphylococcus aureus chronic</p><p>osteomyelitis that closely mimics the human infection: an integrated view of disease pathogen‑</p><p>esis. Am J Pathol. 2012;181(4):1206–1214.</p><p>82. Tuchscherr L, Geraci J, Löffler B. Staphylococcus aureus Regulator Sigma B is Important to</p><p>Develop Chronic Infections in Hematogenous Murine Osteomyelitis Model. Pathogens.</p><p>2017;6(3):31.</p><p>83. Norden CW. Experimental osteomyelitis. IV. Therapeutic trials with rifampin alone and in</p><p>combination with gentamicin, sisomicin, and cephalothin. J Infect Dis. 1975;132(5): 493–499.</p><p>84. Montanaro L, Campoccia D, Arciola CR. Advancements in molecular epidemiology of</p><p>implant infections and future perspectives. Biomaterials. 2007;28(34):5155–5168.</p><p>85. Xue‐Wen H, Jie P, Xian‐Yuan A, et al. Inhibition of bacterial translation and growth by pep‑</p><p>tide nucleic acids targeted to domain II of 23S rRNA. J Pept Sci. 2007;13(4):220–226.</p><p>7 Preclinical Models of Infection in Bone and Joint Surgery 115</p><p>86. Hatamoto M, Ohashi A, Imachi H. Peptide nucleic acids (PNAs) antisense effect to bacterial</p><p>growth and their application potentiality in biotechnology. Appl Microbiol Biotechnol.</p><p>2010;86(2):397–402.</p><p>87. Teiber JF, Horke S, Haines DC, et al. Dominant role of paraoxonases in inactivation of the</p><p>Pseudomonas aeruginosa quorum‐sensing signal N‐(3‐oxododecanoyl)‐L‐homoserine lactone.</p><p>Infect Immun. 2008;76(6):2512–2519.</p><p>88. Eidem TM, Roux CM, Dunman PM. RNA decay: a novel therapeutic target in bacteria. Wiley</p><p>Interdiscip Rev RNA. 2012;3(3):443–454.</p><p>89. McCarthy AJ, Lindsay JA. Genetic variation in Staphylococcus aureus surface and immune</p><p>evasion genes is lineage associated: implications for vaccine design and host‐pathogen interac‑</p><p>tions. BMC Microbiol. 2010;10:173.</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>117</p><p>Chapter 8</p><p>Introduction</p><p>Bacterial and fungal joint infections in children are medical emergencies. If diagnosis and</p><p>treatment are delayed or inadequate, severe morbidity, irreversible joint damage, and even</p><p>fatalities may result. Septic arthritis is more common in childhood than in any other period,</p><p>and more than half of cases are diagnosed in individuals younger than 20 years of age.</p><p>Epidemiology</p><p>The estimated annual incidence of pediatric joint infections in the western world ranges</p><p>between 2 and 10 cases per 100,000 and is much higher in the developing world and</p><p>indigent populations [1,2]. A male‐to‐female ratio >1 has been consistently reported in</p><p>large patients’ series [3].</p><p>Because over 95% of cases of pediatric septic arthritis are of hematogenous origin, the</p><p>age distribution of patients with joint infection is markedly skewed to the left, reflecting</p><p>the increased attack rate of bacteremia in early childhood  [3]. In a large case series</p><p>comprising 725 children, 54% percent of all cases were diagnosed below the age of 2 years,</p><p>25% in children aged 2–5 years, 15% among children aged 6–10 years, and the remaining</p><p>6% in the 11–15‐year‐old group [4]. The delayed maturation of the T‐cell independent arm</p><p>of the immune system in humans results in impaired production of antibodies to bacterial</p><p>polysaccharides below 2–4 years of age  [5]. Thus, this age group has an increased</p><p>susceptibility to encapsulated organisms such as Haemophilus influenzae type b, Kingella</p><p>kingae, or pneumococci [5,6]. On the other hand, the incidence of infectious arthritis in</p><p>infants younger than 6 months is low, indicating vertically acquired immunity and relative</p><p>lack of social contacts, resulting in reduced exposure to potential pathogens in early life.</p><p>Native Joint Arthritis in Children</p><p>Pablo Yagupsky</p><p>118 Bone and Joint Infections</p><p>Microbiology</p><p>Specific Microorganisms and Predisposition</p><p>Because joint infections in children usually result from hematogenous seeding, the</p><p>etiology of septic arthritis frequently overlaps that of pediatric bacteremia. However,</p><p>some microorganisms such as Staphylococcus aureus or K. kingae are remarkably</p><p>overrepresented in childhood arthritis, indicating joint tissue tropism [7].</p><p>As a rule, pediatric septic arthritis is caused by the intra‐articular invasion of a single</p><p>bacterial (or fungal) species. Isolation of multiple organisms should raise the suspicion of</p><p>culture contamination, immunodeficiency, intravenous drug use, or penetrating trauma</p><p>with direct inoculation of microorganisms into the joint space. The patient’s age and</p><p>presence of associated extra‐articular symptoms and signs may provide clues to the likely</p><p>bacterial etiology, as shown in Tables 8.1 and 8.2.</p><p>Staphylococcus aureus is the most common cause of joint infections in neonates, as well</p><p>as in children older than 4 years. This organism is characterized by a wide array of</p><p>virulence factors and genetic determinants of antibiotic resistance. In recent years,</p><p>methicillin‐resistant strains of S. aureus (MRSA) are being increasingly detected in many</p><p>regions, whereas the rate of infection caused by methicillin‐susceptible S. aureus remains</p><p>Table 8.1. Etiology of pediatric hematogenous septic arthritis by age group.</p><p>Age Organism</p><p><2 months Staphylococcus aureus</p><p>Streptococcus agalactiae</p><p>Enterobacteriaceae</p><p>Candida speciesa</p><p>Coagulase‐negative staphylococcia</p><p>Neisseria gonorrhoeae</p><p>2 months to ≤2 years Kingella kingae</p><p>Haemophilus influenzaeb</p><p>Staphylococcus aureus</p><p>Streptococcus pneumoniaeb</p><p>Streptococcus pyogenes</p><p>2–4 years Kingella kingae</p><p>Staphylococcus aureus</p><p>Streptococcus pyogenes</p><p>5–15 years Staphylococcus aureus</p><p>Streptococcus pyogenes</p><p>>15 years Staphylococcus aureus</p><p>Neisseria gonorrhoeaec</p><p>a In premature babies with indwelling vascular catheters.</p><p>b Among unvaccinated and incompletely vaccinated children.</p><p>c In sexually active adolescents.</p><p>8 Native Joint Arthritis in Children 119</p><p>Table 8.2. Etiology of pediatric septic arthritis and associated clinical conditions.</p><p>Associated Condition Possible etiology</p><p>Skin</p><p>Pyoderma Staphylococcus</p><p>aureus, Streptococcus pyogenes</p><p>Varicella lesions Streptococcus pyogenes, Kingella kingae</p><p>Erythematous rash Streptococcus pyogenes</p><p>Erythema migrans Borrelia burgdorferi</p><p>Petechial rash Neisseria meningitidis</p><p>Mucosae</p><p>Gingivostomatitis Kingella kingae</p><p>Urethritis/genital infection Neisseria gonorrhoeae, Ureaplasma spp., Mycoplasma</p><p>hominis</p><p>Respiratory system</p><p>Pneumonia Staphylococcus aureus, Streptococcus pneumoniae,</p><p>Haemophilus influenzae type b</p><p>Necrotizing pneumonia Staphylococcus aureus (especially community‐acquired</p><p>MRSA)</p><p>Cardiovascular</p><p>Endocarditis Staphylococcus aureus, Kingella kingae</p><p>Central nervous system</p><p>Meningitis Streptococcus agalactiae, Streptococcus pneumoniae,</p><p>Haemophilus influenzae type b, Neisseria meningitidis,</p><p>Borrelia burgdorferi</p><p>Hepatosplenomegaly Brucella spp.</p><p>Multifocal involvement Staphylococcus aureus, Haemophilus influenzae type b,</p><p>Neisseria gonorrhoeae, Brucella spp.</p><p>Hemoglobinopathies Salmonella enterica, Streptococcus pneumoniae, other</p><p>Enterobacteriaceae</p><p>Immune system</p><p>HIV infection Staphylococcus aureus, Streptococcus pneumoniae,</p><p>Mycobacterium spp., Nocardia spp., fungi</p><p>X‐linked agammaglobulinemia Encapsulated bacteria, Mycoplasma spp. and</p><p>Ureaplasma spp.</p><p>Common variable immunodeficiency Mycoplasma spp. and Ureaplasma spp.</p><p>Chronic granulomatous disease Staphylococcus aureus, Serratia marcescens,</p><p>Pseudomonas aeruginosa, non‐tuberculous</p><p>mycobacteria, Aspergillus spp.</p><p>Addictions</p><p>Intravenous‐drug addicts Staphylococcus aureus, Pseudomonas aeruginosa</p><p>120 Bone and Joint Infections</p><p>stable [8]. Community‐associated MRSA (CA‐MRSA) infections affect patients lacking</p><p>traditional risk factors for nosocomial MRSA disease; most strains elaborate the Panton‐</p><p>Valentine leukocidin (PVL), that causes lysis of white blood cells, and are generally</p><p>susceptible to antibiotics other than β‐lactams [8–10]. Infections with CA‐MRSA involve</p><p>skin, soft tissues, the lung, and the skeletal system, and are characterized by remarkable</p><p>tissue destruction, a pronounced inflammatory response, and a high incidence of</p><p>complications requiring repeated surgical interventions, intensive care unit admissions,</p><p>longer hospitalization, and results in residual morbidities [10].</p><p>Streptococcus pyogenes (group A Streptococcus) is isolated in 10–20% of preschool and</p><p>early‐school children with septic arthritis. Most cases are detected during the winter and</p><p>spring [11], and it is especially common in patients with concomitant skin infections or</p><p>chickenpox [12]. Despite the exquisite susceptibility of S. pyogenes to β‐lactam antibiotics,</p><p>the disease may be associated with severe sepsis, toxic shock, and mortality [12].</p><p>Streptococcus agalactiae (group B Streptococcus) is diagnosed almost exclusively in the</p><p>neonatal period [13]. Remarkably, S. agalactiae arthritis commonly affects the shoulder.</p><p>In children born in breech presentation, it typically involves the hip joint, suggesting that</p><p>trauma and local hyperemia in the course of bacteremia facilitate seeding of the organism</p><p>into the articular space. The disease usually occurs as a manifestation of late‐onset</p><p>disease (7–90 days after birth). This may explain that current guidelines to detect maternal</p><p>colonization and prevent early transmission of the bacterium to the offspring have not</p><p>substantially reduced its incidence [14].</p><p>Streptococcus pneumoniae arthritis is most common between the ages of six months</p><p>and two years  [14]. In countries, where the conjugate pneumococcal vaccine has been</p><p>introduced, the incidence of invasive S. pneumoniae diseases, including those affecting</p><p>the skeletal system, has substantially decreased [15].</p><p>Haemophilus influenzae type b was the most common cause of septic arthritis in</p><p>children younger than 2 years before the advent of the conjugate vaccine, accounting for</p><p>almost one half of the cases [16]. Children with H. influenzae type b arthritis frequently</p><p>presented with other foci of infection such as meningitis (in 30% of patients),</p><p>osteomyelitis (in 22%), cellulitis (in 30%), pneumonia (in 4%), and otitis media (in</p><p>35%)  [16]. Nowadays, H. influenzae type b has become rare in countries where</p><p>immunization coverage is high [17], but there is a worldwide increase in the incidence of</p><p>invasive disease, including skeletal system infections, caused by H. influenzae type a [18].</p><p>Arthritis due to encapsulated H. influenzae strains other than a or b, non‐encapsulated</p><p>H. influenzae, and Haemophilus species other than influenzae almost exclusively affects</p><p>immunocompromised children [19].</p><p>The increasing use of blood culture vials for seeding skeletal system exudates and</p><p>species‐specific nucleic acid amplification tests (NAATs) has resulted in the recognition</p><p>of K. kingae, a fastidious Gram‐negative member of the normal pharyngeal flora, as the</p><p>most common etiology of septic arthritis and other skeletal system infections below 4</p><p>years of age [7,20,21]. Antecedent or concomitant stomatitis and/or signs of an upper</p><p>respiratory tract infection are frequent, suggesting that invasion of the bloodstream and</p><p>dissemination of the organism to the joints is facilitated by breaching of the mucosal</p><p>layer by a preceding intercurrent viral disease. The disease is characterized by a mild</p><p>systemic and local inflammatory reaction, requiring a high index of suspicion  [21].</p><p>Outbreaks of K. kingae osteoarthritis and other invasive diseases have been reported in</p><p>daycare center facilities [21].</p><p>Although the incidence of arthritis in invasive meningococcal disease is as high as 14%,</p><p>true invasion of the joint by Neisseria meningitidis is uncommon [22]. In most cases, signs</p><p>8 Native Joint Arthritis in Children 121</p><p>of joint inflammation develop several days after initiation of antibiotic therapy and the</p><p>synovial fluid is usually sterile, suggesting an immune‐complex‐mediated phenomenon [22].</p><p>Recurrent disease, a prolonged course, isolation of uncommon N. meningitidis serogroups,</p><p>and family clustering of cases should raise the possibility of complement or properdin</p><p>deficiencies [23].</p><p>Neisseria gonorrhoeae becomes common in sexually active adolescents and its isolation in</p><p>children beyond the neonatal period is definitive proof of sexual abuse. It may be</p><p>pauciarticular and associated with tenosynovitis and skin lesions. Rarely, gonococci may</p><p>infect the joint in the course of disseminated disease in neonates born to infected mothers [24].</p><p>Invasion of the joint space by Salmonella enterica has been reported in children</p><p>suffering from sickle cell anemia and other hemoglobinopathies, and in those living in</p><p>poverty in developing countries. Other Enterobacteriaceae, especially Escherichia coli and</p><p>Klebsiella pneumoniae, are associated with suppurative arthritis in the neonatal period</p><p>and immunocompromised patients [25].</p><p>Pseudomonas aeruginosa is a rare cause of septic arthritis in the general pediatric</p><p>population. However, it may cause joint infection in neonates, in patients with vascular</p><p>catheter‐related infection, immunodeficient children, and intravenous drug‐using</p><p>adolescents [26].</p><p>As the result of effective public health measures, human brucellosis has been eradicated</p><p>from most western countries. In children with arthritis residents of endemic countries</p><p>and travelers returning from these regions, the possibility of brucellar arthritis should be</p><p>considered (see Chapter 19). The disease is characterized by pain, limited mobility, and</p><p>swelling, whereas local redness or warmth is rarely found. Brucellosis usually affects the</p><p>weight‐bearing articulations, especially the hip (in half of the cases). The involvement of</p><p>multiple joints is seen in one‐quarter of patients [27].</p><p>Lyme disease should be included in the differential diagnosis of children exposed to</p><p>ticks in endemic areas who present with arthritis involving large joints (with the</p><p>noticeable exception of the hip) [28]. Migratory arthralgia is present in 18% of children</p><p>with Borrelia burgdorferi infections</p><p>and frank arthritis in 10% [28]. Despite the presence</p><p>of impressive joint inflammation and large effusions, children do not look ill, motion is</p><p>possible, fever is absent in half of the cases, and the white blood cell (WBC) count is</p><p>within normal limits [28].</p><p>Septic arthritis caused by Mycoplasma spp. and Ureaplasma species is almost</p><p>exclusively detected in patients with X‐linked agammaglobulinemia, common variable</p><p>immunodeficiency, or after organ transplantation [29].</p><p>Hematogenous septic arthritis caused by anaerobic organisms is exceptionally seen in</p><p>children and is usually caused by a single bacterial species, generally a gram‐negative bacillus.</p><p>Whenever a penetrating wound or bite is the mechanism of infection, multiple organisms,</p><p>including both aerobes and anaerobes, may be isolated in the joint fluid culture [30].</p><p>Arthritis is the most common manifestation of tuberculosis in the skeletal system after</p><p>Pott disease  [31]. Usually, Mycobacterium tuberculosis bacilli are seeded in synovial</p><p>tissues by the hematogenous route during primary infection. More rarely, the disease</p><p>spreads from a contiguous focus such as the invasion of the atlantoaxial joint from an</p><p>apical pulmonary infection. Tubercular arthritis is monoarticular in 90% of cases and,</p><p>although it can affect virtually any joint, usually involves the hip or knee  [31].</p><p>Constitutional symptoms, such as fever and weight loss, occur in only a minority of</p><p>children. Granulomatous changes and cartilage erosion result in a chronic effusion and</p><p>progressive joint destruction. Signs of acute inflammation are frequently absent, whereas</p><p>local deformity and restricted motion range are typically observed.</p><p>122 Bone and Joint Infections</p><p>Rat‐bite fever is a rare zoonosis caused by two members of rodents’ oral flora</p><p>Streptobacillus moniliformis, mostly in western countries and Australia, and by Spirillum</p><p>minus in Asia [32]. The site of inoculation of the disease usually heals before a septicemic</p><p>disease, frequently characterized by fever and rash, develops. Arthritis involving multiple</p><p>joints is commonly seen in rat‐bite disease caused by S. moniliformis, but is rare in spirillar</p><p>infection. The culture of the synovial fluid exudate is frequently negative, suggesting a</p><p>reactive mechanism [32].</p><p>Candida species and coagulase‐negative staphylococci are pathogens of low‐virulence</p><p>that can cause infectious arthritis in premature babies and neonates in the intensive care</p><p>setting and in young infants with indwelling vascular catheters [33].</p><p>Culture‐Negative Septic Arthritis</p><p>On average, in 33% of children with presumptive joint infections, blood and synovial</p><p>fluid cultures reveal no growth [4,6,7], with percentages ranging between 16% [34] and</p><p>69% [35]. This wide variation reflects differences in the sensitivity of the microbiological</p><p>methods, the wide array of inclusion criteria employed in the different studies, or previous</p><p>administration of antibiotic therapy [36]. Pediatric patients with culture‐negative disease</p><p>consistently show a trend towards younger age, lower body temperature, WBC count,</p><p>and C‐reactive protein (CRP) values on admission, a milder clinical course, a shorter</p><p>hospital stay, and a better prognosis, suggesting that fastidious pathogens of low virulence</p><p>could be responsible for many of these infections [35,37–39].</p><p>The routine inclusion of sensitive NAATs in the workup of young children with</p><p>suspected joint infections, and particularly those targeting specific K. kingae DNA</p><p>sequences, has significantly improved the microbiological diagnosis of septic arthritis</p><p>and confirmed that many culture‐negative cases of septic arthritis are caused by this</p><p>microorganism [20,21]. However, even when sensitive NAATs are used, still one‐fifth of</p><p>cases remain bacteriologically unconfirmed, indicating that many joint infections are</p><p>caused by pathogens that are not detected by the current laboratory methods  [21].</p><p>Recently, metagenomic approaches based on shotgun next‐generation sequencing have</p><p>been employed to identify the etiologic agents of culture‐negative prosthetic joint</p><p>infections with good results [40]. The strategy can detect all the microorganisms that are</p><p>present in a clinical specimen, including pathogens for which target sequences are not</p><p>currently available. Despite these theoretical advantages, many technical problems remain</p><p>to be solved, including the environmental background DNA contamination, the relative</p><p>paucity of pathogen’s DNA in the synovial fluid specimen relative to the abundance of</p><p>host’s sequences, need for extensive databases, and high cost [40]. It is to be expected that</p><p>this and other future culture‐independent diagnostic methods will reduce or even</p><p>eliminate the bacteriologically unconfirmed cases.</p><p>Pathogenesis</p><p>The synovial membrane is highly vascular and lacks a limiting basement membrane,</p><p>enabling easy bacterial access to the joint space in the course of a bacteremic episode.</p><p>Once organisms have penetrated the joint, the low fluid shear conditions facilitate</p><p>microbial adherence  [41]. Uncommonly, pediatric septic arthritis may also result from</p><p>direct inoculation of organisms in the joint by a human or animal bite, joint taps especially</p><p>with the injection of corticosteroids, or surgical procedures.</p><p>8 Native Joint Arthritis in Children 123</p><p>Invasion of the joint space in neonates occurs in the majority of cases as the result of</p><p>the dissemination of the infection from a contiguous metaphyseal focus of</p><p>osteomyelitis  [42]. In young children, the cartilaginous epiphyses receive their blood</p><p>supply from a metaphyseal capillary network that obliterates between 6 and 9 months of</p><p>age. Therefore, infection of a metaphyseal site can easily spread across the growth plate</p><p>to the epiphysis and joint space. Because in older children the epiphyses and metaphyses</p><p>have a separate blood supply and only the metaphyses of the hip, shoulder, and ankle</p><p>bones remain intracapsular, the spread of infection from bone to joints becomes less</p><p>common [42]. Occasionally, the neonatal joint may be directly invaded during a bacteremic</p><p>episode and, in the hospital setting, by direct inoculation of skin organisms during a</p><p>femoral venipuncture [43]. The source of the preceding bacteremia may be the result of</p><p>nosocomial transmission of virulent S. aureus, the newborn’s normal skin flora, or</p><p>acquisition of maternal organisms when delivered through a birth canal colonized by</p><p>Enterobacteriaceae, S. agalactiae, or N. gonorrhoeae [4,12,24].</p><p>Bacteria implicated in septic arthritis usually display a variety of surface‐exposed</p><p>receptors that recognize adhesive matrix molecules, such as collagen, fibronectin, and</p><p>elastin, facilitating invasion by firmly anchoring the organism to the synovial layer [38].</p><p>Local trauma may unveil these tissue components, promoting bacterial adherence, and</p><p>increasing the risk of suppurative arthritis. Inactivation of the genes encoding bacterial</p><p>adhesins significantly reduces the capability of the organisms to establish a joint</p><p>infection [41].</p><p>Both bacterial factors and the host’s immune response contribute to the progressive</p><p>destruction of joint tissues [41]. Bacteria such as S. aureus are internalized by osteoblasts,</p><p>causing apoptosis or evading the immune response by surviving and multiplying in the</p><p>intracellular milieu and releasing potent toxins and enzymes that break down host tissues</p><p>and provide nutrients for bacterial growth. The presence of bacteria in the joint induces</p><p>a strong inflammatory response consisting of proliferation of the synovial cells, leukocyte</p><p>migration, and formation of granulation tissue and abscesses. Synoviocytes and</p><p>infiltrating leukocytes release proteases and secrete cytokines such as interleukin‐1‐β,</p><p>interleukin‐6, and tumor necrosis factor‐α [41]. These cytokines activate an inflammatory</p><p>cascade releasing acute‐phase reactants from the liver, such as CRP, that adhere to</p><p>invading bacteria and facilitate opsonization</p><p>therapies</p><p>are looked for. The use of bacteriophages is a promising option in patients with bone and</p><p>joint infection caused by multiresistant bacteria. Bacteriophages have a long history,</p><p>however it is only recently that experimental and clinical data appeared in the literature</p><p>[16]. In the near future, controlled clinical trials will show their role in biofilm</p><p>infections.</p><p>The role of the bone/serum ratio in the antimicrobial treatment of bone and joint</p><p>infections is still a matter of debate. Important methodological differences have to be</p><p>considered to adequately judge data on bone penetration. These data are often contro-</p><p>versially discussed in the literature, mainly due to the use of varying experimental tech-</p><p>niques in different studies [17,18]. Distinct differences in the extent of bone penetration</p><p>by various classes of antimicrobial agents have been observed. However, the proof for</p><p>the clinical relevance of these differences is still missing. Thus, knowledge about phar-</p><p>macokinetics and pharmacodynamics of antibiotics in bone should stimulate planning</p><p>of clinical studies to fill this missing gap. Many current treatment concepts are based</p><p>on preclinical studies in vitro and in animals [19]. Such data are especially important</p><p>for the management of implant‐associated infections, a field in which controlled clini-</p><p>cal trials are lacking.</p><p>Septic arthritis encompasses a non‐homogenous group of joint infections. In this</p><p>book, eight different clinical situations are covered. In arthritis of children, many aspects</p><p>differ from arthritis in adults. In children, Kingella kingae plays a prominent role, a micro-</p><p>organism which in adults almost exclusively causes endocarditis [20]. In addition,</p><p>Streptococcus agalactiae is still common in neonates. In contrast, Haemophilus influenzae</p><p>type b almost disappeared in young children due to the effective conjugate vaccine.</p><p>Arthritis of axial joints is rare and difficult to diagnose. IV‐drug use is the most frequent</p><p>risk factor of all types of axial arthritis, namely of the sternoclavicular joint, the symphy-</p><p>sis pubis, and the sacroiliac joint. Surgery is rarely needed if the diagnosis is rapidly made</p><p>and the patient has no pyogenic complications. Prosthetic joints are increasingly used not</p><p>only in hip and knee, but also in other joints, mainly shoulder, ankle, and elbow. The</p><p>perioperative infection rate ranges from about 0.5–1.5% after hip or knee arthroplasty up</p><p>to 10% after elbow or ankle joint replacement. Since many aspects vary between the</p><p>different joint prostheses, separate chapters deal with periprosthetic joint infection in</p><p>this book.</p><p>Osteomyelitis encompasses a large spectrum of different diseases. Many different</p><p>classifications are used, depending on different aspects of disease (e.g. pathogenesis,</p><p>duration, presence of implant) and according to the specialist who is managing the case</p><p>(e.g. orthopedic surgeon, infectious disease specialist, pediatrician, angiologist). In this</p><p>book, aspects of age (children, adults), duration of disease (acute, subacute, chronic),</p><p>presence of implant, anatomic location (long bones, vertebrae, jaws), and presence of</p><p>diabetes are presented in separate chapters.</p><p>Together with all authors, I trust that this multidisciplinary book will allow the</p><p>gathering of rapid and exhaustive information regarding all types of bone and joint</p><p>infection. If this book allows you to improve patient management, we have reached</p><p>our goal.</p><p>1 Introduction 3</p><p>References</p><p>1. Torres A, Zhong N, Pachl J, et al. Ceftazidime‐avibactam versus meropenem in nosoco-</p><p>mial pneumonia, including ventilator‐associated pneumonia (REPROVE): a randomised,</p><p>double‐blind, phase 3 non‐inferiority trial. Lancet Infect Dis. 2018;18(3):</p><p>285–295.</p><p>2. Wagenlehner FME, Cloutier DJ, Komirenko AS, et al. Once‐daily plazomicin for complicated</p><p>urinary tract infections. N Engl J Med. 2019;380(8):729–740.</p><p>3. Wirz Y, Meier MA, Bouadma L, et al. Effect of procalcitonin‐guided antibiotic treatment on</p><p>clinical outcomes in intensive care unit patients with infection and sepsis patients: a patient‐</p><p>level meta‐analysis of randomized trials. Crit Care. 2018;22(1):191.</p><p>4. Zimmerli W, Widmer AF, Blatter M, et  al. Role of rifampin for treatment of orthopedic</p><p>implant‐related staphylococcal infections: a randomized controlled trial. Foreign‐Body</p><p>Infection (FBI) Study Group. JAMA. 1998;279(19):1537–1541.</p><p>5. Bernard L, Dinh A, Ghout I, et al. Antibiotic treatment for 6 weeks versus 12 weeks in patients</p><p>with pyogenic vertebral osteomyelitis: an open‐label, non‐inferiority, randomised, controlled</p><p>trial. Lancet. 2015;385(9971):875–882.</p><p>6. Gellert M, Hardt S, Koder K, et al. Biofilm‐active antibiotic treatment improved the outcome</p><p>of knee periprosthetic joint infection: Results from a 6‐year prospective cohort. Int J Antimicrob</p><p>Agents. 2020:105904.</p><p>7. Roux S, Valour F, Karsenty J, et al. Daptomycin > 6 mg/kg/day as salvage therapy in patients</p><p>with complex bone and joint infection: cohort study in a regional reference center. BMC Infect</p><p>Dis. 2016;16:83.</p><p>8. Lowik CAM, Parvizi J, Jutte PC, et al. Debridement, antibiotics and implant retention is a</p><p>viable treatment option for early periprosthetic joint infection presenting more than four weeks</p><p>after index arthroplasty. Clin Infect Dis. 2019.</p><p>9. Depypere M, Morgenstern M, Kuehl R, et al. Pathogenesis and management of fracture‐</p><p>related infection. Clin Microbiol Infect 2019; https://doi.org/10.1016/j.cmi.</p><p>2019.08.006.</p><p>10. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic‐joint infections. N Engl J Med.</p><p>2004;351(16):1645–1654.</p><p>11. Osmon DR, Berbari EF, Berendt AR, et al. Diagnosis and management of prosthetic joint</p><p>infection: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect</p><p>Dis. 2013;56(1):e1–e25.</p><p>12. Berbari EF, Kanj SS, Kowalski TJ, et al. Infectious Diseases Society of America (IDSA) clini-</p><p>cal practice guidelines for the diagnosis and treatment of native vertebral osteomyelitis in</p><p>adults. Clin Infect Dis. 2015;61(6):e26–46.</p><p>13. Lipsky BA, Berendt AR, Cornia PB, et  al. Infectious Diseases Society of America clinical</p><p>practice guideline for the diagnosis and treatment of diabetic foot infections. Clin Infect Dis.</p><p>2012;54(12):e132–173.</p><p>14. Street TL, Sanderson ND, Atkins BL, et al. Molecular diagnosis of orthopedic‐device‐related</p><p>infection directly from sonication fluid by metagenomic sequencing. J Clin Microbiol.</p><p>2017;55(8):2334–2347.</p><p>15. Thoendel MJ, Jeraldo PR, Greenwood‐Quaintance KE, et al. Identification of prosthetic joint</p><p>infection pathogens using a shotgun metagenomics approach. Clin Infect Dis.</p><p>2018;67(9):1333–1338.</p><p>16. Tkhilaishvili T, Winkler T, Muller M, et al. Bacteriophages as adjuvant to antibiotics for the</p><p>treatment of periprosthetic joint infection caused by multidrug‐resistant pseudomonas aerugi-</p><p>nosa. Antimicrob Agents Chemother. 2019;64(1).</p><p>17. Mouton JW, Theuretzbacher U, Craig WA, et  al. Tissue concentrations: do we ever learn?</p><p>J Antimicrob Chemother. 2008;61(2):235–237.</p><p>https://doi.org/10.1016/j.cmi.2019.08.006</p><p>https://doi.org/10.1016/j.cmi.2019.08.006</p><p>4 Bone and Joint Infections</p><p>18. Landersdorfer CB, Bulitta JB, Kinzig M, et al. Penetration of antibacterials into bone: phar-</p><p>macokinetic, pharmacodynamic and bioanalytical considerations. Clin Pharmacokinet.</p><p>2009;48(2):89–124.</p><p>19. Vanvelk N, Morgenstern M, Moriarty TF, et al. Preclinical in vivo models of fracture‐related</p><p>infection: a systematic review and critical appraisal. Eur Cell Mater. 2018;36:184–199.</p><p>20. Yagupsky P. Kingella kingae: from medical rarity to an emerging paediatric pathogen. Lancet</p><p>Infect Dis. 2004;4(6):358–367.</p><p>Bone and Joint Infections: From Microbiology to Diagnostics and Treatment, Second Edition.</p><p>Edited by Werner Zimmerli.</p><p>© 2021 John Wiley & Sons Ltd. Published 2021 by John Wiley & Sons Ltd.</p><p>5</p><p>Chapter 2</p><p>Introduction</p><p>A meticulous diagnostic workup is paramount for a successful management of bone and</p><p>and complement activation. On the other</p><p>hand, cytokines increase the release of host matrix metalloproteinases, such as stromelysin,</p><p>and other collagen‐degrading enzymes. The inflammatory process triggers fluid</p><p>accumulation, increasing intra‐articular pressure and inducing tissue ischemia and</p><p>necrosis [41]. The resulting cartilage destruction causes narrowing of the joint space and</p><p>further erosive damage, leading to disabling orthopedic sequelae.</p><p>Clinical Presentation</p><p>Typically, septic arthritis exhibits a more acute presentation than osteomyelitis, and</p><p>most children with joint infections are brought to medical attention within two to</p><p>five days from the onset of symptoms. Hematogenous pediatric septic arthritis affects</p><p>a single joint in 95% of cases. The involvement of multiple articulations suggests a</p><p>viral, reactive arthropathy or an immunocompromising condition. Polyarticular</p><p>septic arthritis, however, has been noted in neonates, in half of the cases caused by</p><p>gonococci, in 7% of those caused by S. aureus or H. influenzae type b, and in infections</p><p>by Candida species [4].</p><p>124 Bone and Joint Infections</p><p>Septic arthritis usually affects the large weight‐bearing joints of the lower extremities</p><p>(Figure 8.1). Small joints of the hand and feet are overrepresented in K. kingae infections,</p><p>the sacroiliac joints are typically affected in brucellosis, and the sternoclavicular</p><p>articulation by P. aeruginosa in intravenous‐drug users [26] and as a rare complication of</p><p>subclavian vein catheterization [44] (see Chapter 10).</p><p>Most children with septic arthritis present with acute onset of fever and local</p><p>inflammatory changes, such as swelling or localized erythema of the overlying skin.</p><p>Irritability, pain, abnormal (antalgic) posture, restricted range of motion or refusal to</p><p>move the affected extremity or bear weight, and limping are frequent complaints. The</p><p>pain of untreated septic arthritis is continuous and progressive, in contrast to inflammatory</p><p>arthropathies such as juvenile idiopathic arthritis, where symptoms worsen upon rising in</p><p>the morning.</p><p>Infected joints are splinted by muscle contraction to limit motion and reduce pressure</p><p>and the resulting pain. When the hip joint is involved, the extremity is held in flexion,</p><p>external rotation, and abduction, the infected knee or ankle in slight flexion, and the</p><p>shoulder in adduction and internal rotation. While examining the child, it should be kept</p><p>in mind that arthritis of the hip is frequently difficult to localize and patients may present</p><p>with pain referred to the knee or anterior thigh  [4]. Patients with sacroiliitis exhibit a</p><p>positive FABERE (flexion, abduction, external rotation, extension) test. Painful palpation</p><p>of the joint may be also elicited by direct compression of the iliac wing or by digital</p><p>dorsal compression in rectal examination. Newborns and young patients infected with</p><p>Metacarpal</p><p>0.1%</p><p>Elbow</p><p>14.0%</p><p>Sacroiliac</p><p>0.6%</p><p>Hip</p><p>22.2%</p><p>Knee</p><p>39.6%</p><p>Ankle</p><p>13.3%</p><p>Metatarsal</p><p>0.4%</p><p>Interphalangeal</p><p>0.5%</p><p>Wrist</p><p>4.4%</p><p>Shoulder</p><p>4.7%</p><p>Acromioclavicular</p><p>0.1%</p><p>Sternoclavicular</p><p>0.1%</p><p>Figure 8.1. Anatomical distribution of 781 septic joints diagnosed in 725 children.</p><p>8 Native Joint Arthritis in Children 125</p><p>low‐grade virulence pathogens such as K. kingae or Brucella spp. may be afebrile at the</p><p>time of diagnosis, requiring an increased awareness of the possibility of a joint</p><p>infection [20,27]. In neonates, and especially in premature babies, the clinical picture may</p><p>be dominated by nonspecific signs such as poor feeding, vomiting, abdominal distention,</p><p>tachycardia, tachypnea, hypothermia, irritability or apathy, hypotension, poor perfusion,</p><p>and acidosis  [33]. Meticulous physical examination may disclose limited use of an</p><p>extremity or pseudoparalysis, and subtle signs of local inflammation over the affected</p><p>joint, such as discomfort when handled or having the diaper changed, or swelling of the</p><p>buttock, genitalia, thigh, or the entire extremity. In addition to obtaining synovial fluid</p><p>specimens for culture, a complete sepsis workout, including obtaining blood and urine</p><p>cultures and performance of a lumbar puncture, are indicated before administering</p><p>empiric broad‐spectrum antimicrobial therapy.</p><p>Laboratory Investigation</p><p>The key to the diagnosis of bacterial arthritis in children is a high index of clinical</p><p>suspicion. The diagnosis should be confirmed without delay by aspiration of the joint,</p><p>performed with a large‐bore needle (20‐gauge or larger). Although a comprehensive</p><p>microbiological, biochemical, and cytological study of the synovial fluid is usually</p><p>ordered [45], the only definitive proofs of an infectious etiology of the joint inflammation</p><p>are either demonstration of bacteria in the Gram’s‐stain, growth of an unambiguous</p><p>pathogen in culture, or detection of pathogen‐specific DNA sequences by a NAAT. If</p><p>synovial fluid cannot be obtained by close needle aspiration, the procedure should be</p><p>attempted again with imaging guidance, especially for sites that are not easily accessible</p><p>such as the hips, shoulders, or sacroiliac joints [45].</p><p>The current microbiological approach for diagnosing pediatric septic arthritis is</p><p>summarized in Figure  8.2. This laboratory strategy integrates the traditional Gram</p><p>staining and culture on solid media, as well as novel and improved detection methods</p><p>such as inoculation of the synovial fluid aspirate on blood culture vials and the use of</p><p>sensitive NAATs. The algorithm takes into consideration the age of the child, and the</p><p>presence of specific risk factors such as underlying immune deficiency, exposure to</p><p>zoonotic pathogens, human or animal bites, invasive orthopedic procedures, etc.</p><p>Aspiration of an amount of fluid insufficient for an extensive laboratory workup is</p><p>common in young children or when a small joint is drained. In this case, the performance</p><p>of a Gram’s‐stain, real‐time PCR assays with “universal” bacterial primers and K. kingae‐</p><p>specific primers in children aged 6–48 months, and inoculation of a blood culture vial are</p><p>the best options.</p><p>Any cloudy joint effusion should be considered infectious until proven otherwise.</p><p>Although acute rheumatic fever, Reiter’s disease, and juvenile idiopathic arthritis can</p><p>cause a markedly inflammatory synovial fluid, the highest leukocyte counts are seen in</p><p>patients with septic arthritis, usually in the 50,000 to 200,000 cells per mm3 range, of</p><p>which more than 90% are polymorphonuclear leukocytes. A WBC count higher than</p><p>50,000 leukocytes per mm3 is generally proposed as a cutoff to differentiate septic arthritis</p><p>from non‐infectious joint exudates. Yet lower counts may be seen in infections caused by</p><p>Gram‐negative organisms such as N. gonorrhoeae, K. kingae, and Brucella species, early</p><p>in the course of bacterial arthritis of any etiology, and in neutropenic patients [46,47].</p><p>Conversely, WBC counts >50,000/mm3 of synovial fluid may be observed in children</p><p>126 Bone and Joint Infections</p><p>with juvenile idiopathic arthritis, serum sickness, or reactive arthritis. Measurements of</p><p>the synovial fluid glucose, protein, or lactate contents are neither sensitive nor specific for</p><p>bacterial arthritis [48].</p><p>The aspirate should be transported to the microbiology laboratory without delay in the</p><p>original syringe or a sterile tube. The use of swabs, although inexpensive and easy to use,</p><p>should be discouraged. They are more likely to be contaminated; certain fibers, such as</p><p>cotton, may inhibit bacterial growth; and organisms may remain adherent to swabs resulting</p><p>in a false‐negative Gram’s‐stain examination and reducing the culture’s sensitivity [46].</p><p>A Gram’s stain should be prepared from a centrifuged synovial specimen and carefully</p><p>examined. The test is positive in 75% of patients with staphylococcal arthritis but in less</p><p>than half of those infected by Gram‐negative organisms  [45], probably because of a</p><p>lower bacterial load and the difficulties in recognizing the presence of bacteria against the</p><p>joint infections, as the surgical and medical treatment differ considerably between aseptic</p><p>and septic entities. Since low‐virulent pathogens present with only subtle clinical signs and</p><p>symptoms of infection, the diagnosis represents a challenge. It requires a comprehensive</p><p>approach involving several diagnostic tests, especially in the presence of an implant. Once</p><p>the infection is diagnosed, additional infection features determine the treatment strategy,</p><p>including duration (acuity) of infection, source of infection (pathogenesis), and expected</p><p>or isolated pathogen(s).</p><p>Precise history taking, clinical examination, and imaging help to determine the</p><p>duration of infection. The acuity of infection is particularly relevant in implant‐associated</p><p>infections, as the biofilm age (“maturity”) guides the surgical management. In acute</p><p>infections, the implant can generally be successfully retained, whereas in chronic</p><p>infections, it should be removed or exchanged. Similarly, the duration of infection directs</p><p>the surgical strategy in osteomyelitis. While in acute osteomyelitis without implant,</p><p>surgery is generally dispensable, in chronic osteomyelitis surgical debridement with</p><p>removal of all dead material (sequestrectomy) is required.</p><p>Clues for the pathogen are the acuity of symptoms, pathogenesis, previous episodes of</p><p>bone or joint infection, and the local epidemiology [1]. Because source control is required</p><p>to improve the treatment outcome, identification of the hematogenous route of infection</p><p>is of the utmost importance and allows to extend diagnostic measures in order to search</p><p>and identify the primary focus of infection (Table 2.1).</p><p>Depending on the primary focus of infection, the treatment strategy needs to be</p><p>adapted, i.e. the intravenous treatment prolonged to four to six weeks, or the primary</p><p>infection source controlled by surgical intervention. If the primary infection focus is not</p><p>Diagnostic Approach in Bone</p><p>and Joint Infections</p><p>Nora Renz, Donara Margaryan, and Andrej Trampuz</p><p>6 Bone and Joint Infections</p><p>identified and treated accordingly, a secondary infection may reoccur due to relapsing</p><p>spread from the primary focus, despite correct treatment of the bone or joint</p><p>infection [2].</p><p>The preoperative diagnostic methods are generally less sensitive and specific than</p><p>intraoperative methods. Therefore, preoperative diagnostic tests, such as imaging</p><p>Table 2.1. Investigation of cause in case of hematogenous bone and joint infections.</p><p>Pathogen Focus Investigation</p><p>Staphylo‐</p><p>coccus spp.</p><p>Staphylococcus</p><p>aureus</p><p>• Skin</p><p>• Infective endocarditis</p><p>• Intravascular devices</p><p>or catheters</p><p>• Primary bacteremia</p><p>• Skin examination</p><p>• Check for intravascular</p><p>implants</p><p>• Blood cultures (BC)</p><p>• Transesophageal</p><p>echocardiography (TEE)</p><p>in case of positive BC</p><p>Coagulase‐</p><p>negative</p><p>staphylococci</p><p>• Intravascular devices</p><p>or catheters</p><p>• Infective endocarditis</p><p>• Check for intravascular</p><p>implants</p><p>• Blood cultures</p><p>• TEE in case of positive</p><p>BC</p><p>Streptococcus</p><p>spp.</p><p>Viridans group</p><p>(Streptococcus</p><p>mitis/oralis)</p><p>• Oral cavity</p><p>• Infective endocarditis</p><p>• Check for recent dental</p><p>procedure</p><p>• Orthopantomogram</p><p>• Blood cultures</p><p>• TEE in case of positive</p><p>BC</p><p>S. agalactiae,</p><p>S. dysgalactiae</p><p>• Abdomen</p><p>• Urogenital tract</p><p>• Skin</p><p>• Skin examination</p><p>• Urinalysis and culture</p><p>• Imaging of abdomen/</p><p>pelvis in case urinalysis is</p><p>normal</p><p>S. gallolyticus • Colorectal cancer or</p><p>adenoma</p><p>• Infective endocarditis</p><p>• Blood cultures</p><p>• TEE in case of positive</p><p>BC</p><p>• Colonoscopy (if not done</p><p>recently)</p><p>Gram‐</p><p>negative rods</p><p>Escherichia coli,</p><p>Klebsiella spp.,</p><p>Enterobacter spp.,</p><p>Pseudomonas spp.</p><p>• Abdomen</p><p>• Urogenital tract</p><p>• Urinalysis and culture</p><p>• Imaging of abdomen/</p><p>pelvis (colonoscopy) if</p><p>not performed recently</p><p>and urinalysis is normal</p><p>Enterococcus</p><p>spp.</p><p>E. faecalis,</p><p>E. faecium,</p><p>other spp.</p><p>• Abdomen</p><p>• Urogenital tract</p><p>• Infective endocarditis</p><p>• Urinalysis and culture</p><p>• Blood cultures</p><p>• TEE in case of positive</p><p>blood cultures</p><p>• Colonoscopy if not</p><p>performed recently and</p><p>urinalysis is normal</p><p>2 Diagnostic Approach in Bone and Joint Infections 7</p><p>and synovial fluid analysis, need to be re‐evaluated and interpreted together with</p><p>intraoperative findings. Microbiological and histopathological analysis of tissue</p><p>samples, as well as sonication fluid cultures of removed foreign bodies, may allow</p><p>improvement of the etiologic diagnosis. In unclear cases in the preoperative setting,</p><p>one must plan the treatment presuming infection and avoid insufficient surgical</p><p>debridement or partial retention of implant material.</p><p>Common Microorganisms Causing Bone and Joint Infection</p><p>Most pathogens causing bone and joint infections originate from the skin or from</p><p>mucosal surfaces of the oral cavity, urogenital, and intestinal tract, i.e. from the</p><p>patient´s microbiome. The majority of bone and joint infections is caused by Gram‐</p><p>positive cocci, mostly by Staphylococcus spp., which account for 50–60% of all infec-</p><p>tions [3]. The remaining percentages are comprised of Streptococcus spp., Enterococcus</p><p>spp., Gram‐positive and Gram‐negative anaerobes, and other rare pathogens such as</p><p>fungi, mycobacteria, or intracellular bacteria  [4]. After open fractures with dirty</p><p>wounds, environmental microorganisms are involved, whereas in unintentional use of</p><p>contaminated medical devices (breach of sterility), typically nosocomial pathogens are</p><p>observed.</p><p>The proportion of individual pathogens depends on the pathogenesis of infection, the</p><p>anatomic location, and the time of occurrence in case of postoperative infections  [3].</p><p>Whereas native septic arthritis, vertebral osteomyelitis without previous instrumentation,</p><p>and acute osteomyelitis in children are predominantly caused by hematogenous seeding into</p><p>the affected bone or joint, only approximately 30% of periprosthetic joint infections (PJI) are</p><p>caused hematogenously. In infections associated with implants, such as fracture‐fixation</p><p>devices or spine instrumentation hardware, hematogenous infection is extremely rare</p><p>(<10%) [5,6].</p><p>Hematogenous bone and joint infections are mainly caused by highly virulent patho-</p><p>gens and are rarely polymicrobial. Unless antimicrobial treatment is administered</p><p>before diagnostic sampling, false negative cultures are the exception, because of the</p><p>high virulence and the possibility of additional culturing specimens from the primary</p><p>source. In a recent analysis of 106 hematogenous PJI, 41% of episodes were caused by</p><p>S. aureus, 30% by Streptococcus species, 12% by Enterococcus species, and 8% by Gram‐</p><p>negative rods. The most common primary foci were located in the cardiovascular sys-</p><p>tem (31%), skin and soft tissue (22%), oral cavity (22%), and urogenital (17%) or</p><p>gastrointestinal tract (10%); in 32% the primary focus was not identified [2]. These pro-</p><p>portions apply to other bone and joint infections, which are predominantly caused by</p><p>hematogenous seeding, such as vertebral osteomyelitis and native joint arthritis. Bone</p><p>and joint infections caused by low‐virulent pathogens such as coagulase‐negative</p><p>staphylococci and Cutibacterium spp. are rarely hematogenously acquired, and if so, an</p><p>intravascular infection should be assumed (e.g. prosthetic valve endocarditis, intravas-</p><p>cular catheter‐ or cardiac implantable electronic device associated infections). A per-</p><p>sisting or relapsing bacteremia over a long period is required to seed in native bone or</p><p>joint structures. Low‐virulent pathogens typically cause implant‐associated infections,</p><p>because their survival in the human body requires a foreign surface such as an implant</p><p>or a sequestrum [7,8].</p><p>8 Bone and Joint Infections</p><p>Exogenous infections occur perioperatively, i.e. during surgery or in the first days</p><p>thereafter in the case of wound healing disturbances and persistent wound drainage.</p><p>Pathogens from the skin microbiome are mainly involved in perioperative infections. In</p><p>instrumented bones and joints in proximity to the intestinal or urinary tract openings</p><p>(e.g. hip and lower spine segments), Gram‐negative rods are more frequent than in other</p><p>parts of the body, especially in obese and incontinent patients [3,9].</p><p>Polymicrobial infections mainly occur in patients with surgical wound healing</p><p>disturbance and in obese patients  [9]. S. aureus as a member of the non‐resident skin</p><p>microbiome can cause postoperative infections, because of transient skin colonization in</p><p>about one third of healthy human beings. However, whereas in hematogenous infections</p><p>S. aureus is more frequent than coagulase‐negative staphylococci (28% vs. 12%), in</p><p>postoperative infections the opposite is true (14% vs. 41%)  [1]. Table  2.2 shows the</p><p>proportions of hematogenous and perioperatively acquired (exogenous) bone and joint</p><p>infections [1,2,6,10,11].</p><p>In case of 2° or 3° grade open fractures, colonization of the exposed tissue due to</p><p>contact with soil may occur and proceed to infection, if surgical debridement and/or</p><p>antimicrobial prophylaxis are inadequate [12]. In these situations, polymicrobial infec-</p><p>tions and isolation of environmental pathogens such as Gram‐negative bacteria and</p><p>Bacillus spp. are more common  [12,13]. Rarely, exogenous infections occur after</p><p>arthrocentesis or injections/infiltrations, or after spontaneous or traumatic skin</p><p>perforation.</p><p>Diagnostic Approach in Spinal Infection</p><p>In spinal infections, the pathogenesis, diagnostic procedure, and therapy differ whether</p><p>an implant is in place (i.e. spondylodesis‐associated infections) or not (i.e. native</p><p>vertebral osteomyelitis). Figure 2.1 shows a diagnostic algorithm for the evaluation of</p><p>spinal infections.</p><p>Table 2.2. Proportions of causing pathogens stratified according to the route of infection.</p><p>Pathogen</p><p>Hematogenous infection</p><p>[1,2,10,11] Postoperative exogenous infection</p><p>Staphylococcus aureus 39% (28–46%) 27% (14–43%)</p><p>Coagulase‐negative staphylococci 8% (4–13%) 38% (29–48%)</p><p>Streptococcus spp. 26% (12–39%) 3% (1–6%)</p><p>Enterococcus spp. 4% (1–12%) 5% (2–10%)</p><p>Gram‐negative rods 15% (8–21%) 11% (8–22%)</p><p>Cutibacterium spp. 0% (0–1%) 13% (10–19%)</p><p>Other 3% (1–6%) 4% (1–7%)</p><p>Culture‐negative 2% (1–5%) 10% (1–17%)</p><p>Polymicrobial 3% (1–9%) 18% (15–24%)</p><p>Sources: [1,2,6,10,11].</p><p>Unpublished data: [1,6].</p><p>2 Diagnostic Approach in Bone and Joint Infections 9</p><p>New onset back pain or neurologic deficits, accompanied by systemic inflammatory</p><p>signs and/or wound healing disturbances, should prompt a diagnostic assessment. Before</p><p>spine surgery, history of fever/rigors, clinical examination, determination of systemic</p><p>inflammatory markers, and imaging of the spine should be performed (conventional</p><p>Clinical examination</p><p>Laboratory testing (CRP)</p><p>Plain X-ray/CT/MRT</p><p>Septic patient? yes</p><p>yes</p><p>yes</p><p>yes</p><p>yes</p><p>no</p><p>no</p><p>no</p><p>no</p><p>Open or CT-guided biopsy of</p><p>the spine/disc1</p><p>Infection without prior</p><p>spinal intervention?</p><p>Spinal implant in situ?</p><p>R</p><p>ep</p><p>ea</p><p>t b</p><p>io</p><p>ps</p><p>y</p><p>Workup after BC sampling</p><p>BC positive</p><p>TEE</p><p>(vegetation?)</p><p>OPG</p><p>Urinalysis</p><p>X-ray of lung</p><p>Intravascular</p><p>device (vascular</p><p>catheter, CIED,</p><p>prosthetic valve)</p><p>Hematogenous</p><p>infection: search</p><p>for distant focus:</p><p>Contiguous</p><p>infection: search</p><p>for adjacent focus:</p><p>Microbiology</p><p>Histopathology</p><p>+/– sonication</p><p>Surgery with intra-</p><p>operative</p><p>diagnostics3</p><p>BC negative</p><p>Examination of</p><p>skin (cellulitis/</p><p>ulcer?)</p><p>Imaging of</p><p>abdomen/pelvis</p><p>(abscess?)</p><p>Collection of blood cultures</p><p>Neurological deficit due to</p><p>compression?</p><p>Histopathology</p><p>Microbiology</p><p>Conservative treatment with</p><p>antimicrobial therapy</p><p>Histopathology and/or</p><p>microbiology2 consistent with</p><p>infection?</p><p>No clinical improvement/</p><p>progressive mechanical</p><p>instability?</p><p>1 If blood cultures grow the causing pathogen in clinically and radiologically proven vertebral osteomyelitis</p><p>(without implant), biopsy of spine or disc may be omitted.</p><p>2 For highly virulent organisms (e.g. Staphylococcus aureus, Escherichia coli, Streptococcus spp.) or for patients</p><p>on antibiotic therapy one positive sample confirms infection, for low-virulent organisms (e.g. Staphylococcus</p><p>epidermidis, Cutibacterium acnes) ≥2 positive samples are required to confrim infection.</p><p>3 Additional microbiological investigations (Mycobacterium spp., Brucella spp.), if exposure/risk factors present.</p><p>BC: blood cultures, TEE: transesophageal echocardiography, OPG: orthopantomogram, CIED: cardiac</p><p>implantable electronic device</p><p>Spinal instability/deformity?</p><p>Epidural or paravertebral</p><p>abscess (>1 cm)?</p><p>Figure 2.1. Diagnostic algorithm for spinal infections (with and without implant).</p><p>10 Bone and Joint Infections</p><p>X‐ray, CT, or MRI, as appropriate [see Chapter 24]). In case of new onset of back pain</p><p>without prior intervention (i.e. surgery or infiltration/injections) or in febrile patients,</p><p>blood cultures should be collected before antimicrobial therapy. Depending on the</p><p>isolated pathogen, further investigation of the primary source of infection should be</p><p>initiated in parallel to the investigation of the spine.</p><p>Native spine infections are mostly of hematogenous origin and require collection of</p><p>blood cultures. Transcutaneous tissue sampling (in general CT‐guided biopsy) is needed,</p><p>if blood cultures remain negative and conditions requiring surgery are absent. Most cases</p><p>with hematogenous infection can be treated with antibiotics alone [14] (see Chapter 18).</p><p>Therefore, comprehensive microbiological sampling before initiation of antimicrobial</p><p>treatment is crucial. In contrast, implant‐associated spinal infections predominantly</p><p>occur after perioperative contamination. Surgical debridement is always needed, and the</p><p>diagnosis is confirmed with intraoperative diagnostic tests (see Chapter 24).</p><p>During surgery, multiple tissue samples (three to five samples for microbiology and his-</p><p>topathology) are collected and removed hardware (either entire internal fixation material or</p><p>only loose implants) should be sent for sonication fluid culture. Carlson et al. [15] showed</p><p>sonication of removed spinal implants to be more sensitive than peri‐implant tissue culture,</p><p>reaching statistical significance when using a threshold of ≥20 CFU/10mL for sonication</p><p>culture positivity. The removed implant should be sent to sonication to exclude low‐grade</p><p>infection in all spinal surgeries performed for presumed mechanical (aseptic) complications</p><p>such as implant loosening, which may manifest several years after instrumentation. This</p><p>observation is supported by Prinz et al. [16] who demonstrated microbial growth in 22 of</p><p>82 patients (41%) with screw loosening, as compared to none of 28 patients with firm</p><p>screws (p <0.01). Features of native spine infection requiring surgery are mechanical insta-</p><p>bility of the spine due to destruction of the involved vertebral bodies, neurological deficits</p><p>due to compression or epidural, or paravertebral abscesses greater than 1 cm [17]. In these</p><p>cases, immediate surgery and decompression with intraoperative tissue sampling for histo-</p><p>pathological and microbiological analysis should be performed. In all other cases treated</p><p>conservatively, percutaneous or open biopsy of the affected tissue represents an important</p><p>step in the diagnostic algorithm, if the pathogen is not isolated in blood cultures. Antibiotic</p><p>treatment should be withheld in stable patients until the diagnosis of infection is confirmed</p><p>by microbiology or histopathology. If the biopsy results are inconclusive, the diagnostic</p><p>sampling intervention should be repeated [18]. If specific exposures or risks are present,</p><p>additional investigations to detect mycobacteria (acid‐fast stain, PCR, prolonged cultures)</p><p>and other rare pathogens (such as Brucella spp.) are advocated (see Chapter 19).</p><p>Diagnostic Approach in Bone Fixation Device‐Associated Infection</p><p>The aim of the diagnostic workup in case of non‐union, pain, or local wound distur-</p><p>bances after bone fixation is to look for confirmatory signs for infection either preop-</p><p>eratively or during explorative surgery [19,20]. Figure 2.2 shows an algorithm which</p><p>guides through the diagnostic workup. Medical history,</p><p>clinical exam, laboratory</p><p>analysis, and imaging represent the cornerstones of preoperative diagnostics in sus-</p><p>pected bone fixation device‐associated infections [21,22]. Hematogenous route is rare</p><p>in bone fixation device‐associated infections, however secondary bacteremia may</p><p>occur. If the patient is febrile or shows signs indicative for sepsis, blood cultures should</p><p>be collected before initiation of antimicrobial treatment.</p><p>2 Diagnostic Approach in Bone and Joint Infections 11</p><p>Some features confirm the diagnosis of infection (confirmatory criteria), while others</p><p>are only clues of infection, but could be also caused by non‐infectious reasons (suggestive</p><p>criteria)  [20]. Presence of pus or wound discharge, an exposed implant, or sinus tract</p><p>confirm infection and should prompt revision surgery, during which diagnostics with</p><p>Clinical examination</p><p>Laboratory testing (CRP)</p><p>Plain X-ray/CT/MRT</p><p>Septic patient?</p><p>Sinus tract</p><p>Wound dehiscence</p><p>Exposed implant</p><p>Purulent secretion</p><p>no</p><p>no</p><p>no</p><p>≥ 1 confirmatory sign present?</p><p>≥ 1 Suggestive result: ≥ 1 Confirmatory result:</p><p>Suggestive criteria present?</p><p>yes</p><p>yes</p><p>Collection of blood cultures</p><p>Surgery with intraoperative diagnostics</p><p>yes</p><p>Radiological signs1</p><p>Increased CRP</p><p>New joint effusion</p><p>Local wound erythema</p><p>Observation, if surgery for</p><p>mechanical reasons, look for</p><p>confirmatory signs</p><p>Non-significant</p><p>microbiological result2</p><p>Significant</p><p>microbiological result4</p><p>Inconclusive</p><p>histopathology3</p><p>1 Infectious callus, sequestrum, osteolysis, implant loosening, non-union, cortical sclerosis</p><p>2 Single positive tissue culture or sonication culture <50 CFU/ml with low-virulent organisms (e.g. S. epidermidis, C.</p><p>acnes)</p><p>3 Inconclusive: 0–5 neutrophils/high power field (HPF); positive: >5 neutrophils/HPF (at 400x magnification)</p><p>4 ≥1 positive tissue culture or sonication culture with highly virulent organisms (e.g. Staphylococcus aureus,</p><p>Escherichia coli, Streptococcus spp.) or in patients on antibiotic therapy, ≥2 positive samples with low-virulent</p><p>organisms (e.g. Staphylococcus epidermidis, Cutibacterium acnes)</p><p>Infection possible Infection confirmed</p><p>Positive</p><p>histopathology3</p><p>Persistent or increasing</p><p>wound discharge</p><p>Histopathology</p><p>Tissue culture</p><p>Synovial fluid culture</p><p>Sonication</p><p>Figure 2.2. Diagnostic algorithm for bone fixation device‐associated infections.</p><p>12 Bone and Joint Infections</p><p>sampling for microbiological and histopathological examinations is completed. If</p><p>suggestive criteria are present, such as radiological signs (i.e. infectious callus, non‐union,</p><p>implant loosening, cortical sclerosis), increased serum C‐reactive protein (CRP), persis-</p><p>tent wound discharge, or local signs of inflammation (i.e. erythema, excess heat, swelling,</p><p>new effusion of adjacent joint), the threshold for revision surgery should be low. In the</p><p>absence of suggestive criteria, observation is feasible. However, if surgery is performed</p><p>for mechanical reasons, comprehensive diagnostics should be performed intraoperatively</p><p>to exclude a concomitant low‐grade infection.</p><p>The “gold standard” for the diagnosis of infection is a combination of intraoperatively</p><p>collected peri‐implant tissue for culture and histopathological analysis, as well as culture of</p><p>sonication fluid of removed implants and cortical bone fragments (see Chapter  23).</p><p>Analyzing 64 non‐unions after fracture, histopathology with detection of more than 5 neu-</p><p>trophils per high power field at 400x magnification showed a sensitivity of 80% and a speci-</p><p>ficity of 100% for detection of infection [23]. For microbiological analysis of tissue, multiple</p><p>specimens (at least three, not superficial nor sinus tract swabs) should be taken, each with</p><p>clean instruments. In case of effusion in a joint adjacent to a fractured bone, a synovial fluid</p><p>specimen obtained by arthrocentesis under aseptic conditions may be included as a single</p><p>sample [20]. Phenotypically indistinguishable pathogens identified by culture from at least</p><p>two separate deep tissue/implant specimens are considered confirmative in case of low‐</p><p>virulent pathogens such as coagulase‐negative staphylococci, Cutibacterium spp., or</p><p>Corynebacterium spp. Growth of a low‐virulent pathogen in one single specimen culture</p><p>should be considered a suggestive criterion only. In contrast, for highly virulent pathogens</p><p>such as S. aureus, Gram‐negative rods, or streptococci, growth in one single tissue sample</p><p>is sufficient for the diagnosis of infection, as contamination with these pathogens is</p><p>extremely rare. In addition, detection of bacteria or fungi in histopathological examination</p><p>using specific staining techniques confirms the infection, as do pathognomonic histology</p><p>features (e.g. granulomatous inflammation pattern in infection caused by Mycobacterium</p><p>spp. or Brucella spp.). Future research is required to elaborate the diagnostic value of novel</p><p>biomarkers, molecular methods, and nuclear imaging.</p><p>In the presence of at least one confirmative intraoperative result (i.e. significant micro-</p><p>bial growth and/or positive histopathology), infection is confirmed, and respective anti-</p><p>microbial treatment is required (see Chapter 12). However, if no confirmative criterion is</p><p>met after surgery, the presence of one suggestive criterion and even more of several</p><p>suggestive criteria should raise a high suspicion of infection, and respective treatment is</p><p>advocated.</p><p>Diagnostic Approach in Native Arthritis and Infections after</p><p>Anterior Cruciate Ligament Repair (ACL‐R)</p><p>Every patient with acute onset of joint pain or local inflammatory signs without pre-</p><p>ceding trauma should be assessed for and treated as septic arthritis until proven oth-</p><p>erwise. A  distinctive history includes intra‐articular steroid injections, preceding</p><p>infections of the respiratory or gastrointestinal tract, and promiscuity (indicative for</p><p>gonococcal or chlamydial arthritis), as well as signs or symptoms of bacteremia in</p><p>the preceding days/weeks. Differential diagnoses include rheumatic disorders (rheu-</p><p>matoid arthritis, crystallopathy), activated osteoarthritis, and reactive arthritis.</p><p>2 Diagnostic Approach in Bone and Joint Infections 13</p><p>Figure 2.3 shows a diagnostic algorithm. Systemic inflammatory parameters should be</p><p>determined, including differential white blood cell (WBC) count and serum C‐reactive</p><p>protein (CRP). Imaging is rarely necessary in the acute management of septic arthri-</p><p>tis. Conventional x‐ray detects preexisting joint diseases (i.e., osteoarthritis, rheuma-</p><p>toid arthritis, osteomyelitis, or chondrocalcinosis). Ultrasound may be useful to guide</p><p>Immediate collection of blood</p><p>cultures and arthrocentesis,</p><p>then start empiric antibiotics</p><p>Arthrocentesis for:</p><p>Figure 2.3. Diagnostic algorithm for native septic arthritis and infections after anterior cruciate</p><p>ligament reconstruction (ACL-R).</p><p>14 Bone and Joint Infections</p><p>joint aspiration [24]. If infection after ACL‐R is suspected, MRI and CT may be help-</p><p>ful to exclude other differential diagnoses such as graft impingement, rupture, focal</p><p>arthrofibrosis, infected baker cysts, or cystic degeneration of the graft. In addition, it</p><p>should be performed in chronic infections to assess for sequestra. The sensitivity of a</p><p>WBC count >10,000 cells/μL is 90% and a CRP >100 mg/L is 77%, but both param-</p><p>eters are nonspecific at this threshold [24]. If the patient is septic or suffers from fever</p><p>or rigors, immediate collection of blood cultures and arthrocentesis followed by initia-</p><p>tion of empiric treatment is paramount.</p><p>Septic arthritis is a medical emergency due to the rapid cartilage damage by</p><p>leukocyte proteases. Therefore, prompt arthrocentesis  –  preferably before antibiotic</p><p>treatment  –  is the first step in the assessment of a painful inflamed joint, even in</p><p>patients without systemic signs of infection. Synovial fluid examination consists of</p><p>analysis of leukocyte count and percentage of granulocytes (use EDTA tubes to pre-</p><p>vent clumping), Gram stain, conventional culture (inoculate</p><p>preferentially pediatric</p><p>blood culture bottles to increase culture sensitivity), and crystals by polarized light</p><p>microscopy (for details see Chapter 9). Gram stain has a low sensitivity. However, due</p><p>to the high specificity, it can be used as a rule‐in test in case of a positive result. The</p><p>presence of crystals in synovial fluid does not rule out infection, as the combination</p><p>of crystal-induced and septic arthritis may occur [25]. If any of the analyses in syno-</p><p>vial fluid is consistent with infection, immediate surgical intervention (arthroscopy or</p><p>arthrotomy) is indicated.</p><p>Intraoperatively, the diagnosis is completed with histopathological and microbiologi-</p><p>cal analysis of synovial membrane. Furthermore, the patient should be examined for</p><p>primary infection foci, including intravascular, skin and soft‐tissue, urogenital, gastroin-</p><p>testinal, or pulmonary infections, if no intervention or adjacent infectious focus at the</p><p>site of the joint preceded the acute onset of symptoms.</p><p>Novel biomarkers in synovial fluid such as (D‐lactate, interleukin‐6 (IL‐6), total lactate</p><p>D‐ and L‐isomers), and calprotectin have been investigated for the diagnosis of septic</p><p>arthritis in recent years. Among them, mainly D‐lactate and calprotectin showed a prom-</p><p>ising performance (sensitivity of 85% and 76%, respectively, and specificity of 96 and</p><p>94%, respectively) [26,27]. Interleukin‐6, however, did not allow for a reliable differentia-</p><p>tion between septic and aseptic arthritis [28].</p><p>Infections after ACL‐R represent a combination of a native joint arthritis with trans-</p><p>plant or implants in situ (avascular graft and fixation devices). In most cases, concomi-</p><p>tant osteomyelitis of the adjacent bone, where the new graft is attached with fixation</p><p>devices, is present. The most common pathogenesis is perioperative contamination of</p><p>the graft or joint. Most infections occur in the first few weeks after reconstruction and</p><p>present with acute onset of signs and symptoms of infection [29]. Clinical distinction</p><p>between physiological postoperative process and signs of infection is difficult. Similarly,</p><p>the interpretation of synovial fluid leukocyte count is challenging and cut‐off values for</p><p>diagnosing infection in this setting have not been determined yet. This is especially true</p><p>in chronic low‐grade infections caused by low‐virulent pathogens, evoking little local or</p><p>systemic inflammation. Nevertheless, synovial leukocyte count of >10,000/μL may</p><p>indicate septic arthritis, whereas normal counts exclude it. In case of instable transplant,</p><p>exchange of graft/fixation device is required, and the removed parts should be used for</p><p>microbiological analysis, i.e. conventional culture of the graft and sonication of the fixa-</p><p>tion devices [29].</p><p>2 Diagnostic Approach in Bone and Joint Infections 15</p><p>Diagnostic Approach in Periprosthetic Joint Infections (PJI)</p><p>The diagnosis of acute PJI is straightforward, because the clinical context is often</p><p>obvious, and most diagnostic tests have a high sensitivity. In contrast, chronic low‐grade</p><p>infections are difficult to differentiate from aseptic prosthetic failures. Therefore, a com-</p><p>prehensive algorithmic approach combining preoperative and intraoperative results is</p><p>required to reliably confirm or exclude infection (Figure 2.4).</p><p>Every painful prosthetic joint should be assessed for infection. Initial examinations include</p><p>clinical examination of the patient, assessment of systemic inflammatory parameters, and imag-</p><p>ing. Systemic inflammatory parameters are neither sensitive nor specific for the diagnosis of PJI</p><p>(see Chapter 11). Nevertheless, they represent a relevant puzzle piece in the diagnostic workup.</p><p>As in native joint infection, patients presenting with sepsis or with fever/rigors should</p><p>be promptly assessed for hematogenous infection, and blood and synovial fluid should be</p><p>collected, followed by immediate empiric antimicrobial treatment. Revision surgery</p><p>should take place within a few hours after admission in order to realize source control.</p><p>However, septic revision of a prosthetic joint is a challenging procedure requiring special</p><p>expertise. Therefore, it should be performed by an experienced orthopedic surgeon. The</p><p>focus lies on intraoperative diagnostics such as histopathological and microbiological</p><p>analysis including synovial fluid analysis (culture and leukocyte count), periprosthetic</p><p>tissue (at least three to five samples) and sonication of the retrieved implant parts [30]</p><p>(see Chapter 11). Depending on the pathogen and the clinical presentation, investigation</p><p>of the primary focus with additional diagnostic tests should be performed. In late acute</p><p>onset infections with negative blood cultures, the possibility of contiguous infection by</p><p>direct expansion of a nearby focus onto the joint should be considered.</p><p>In chronic infection, the initial focus lies on the preoperative assessment. The infection</p><p>should ideally be diagnosed or excluded before revision, which allows the planning of the</p><p>most appropriate treatment strategy. In the preoperative setting, purulent wound secretion</p><p>and/or sinus tract communicating with the prosthetic joint are confirmative signs of PJI.</p><p>In this case no further diagnostic measures are needed, and revision surgery with intraop-</p><p>erative diagnostics should be scheduled. In the absence of confirmatory clinical signs, the</p><p>most important and sensitive diagnostic measure in the preoperative setting is arthrocen-</p><p>tesis. It should be performed according to standard aseptic technique, preferably in the</p><p>operating room. Culture of synovial fluid has a low sensitivity, as only planktonic bacteria</p><p>are detected, and bacteria embedded in the biofilm on the implant surface remain unrec-</p><p>ognized. Synovial fluid leukocyte count analysis is more sensitive, as it reflects the host</p><p>reaction against the microorganisms. However, in situations associated with aseptic</p><p>inflammatory changes of the joint, the specificity is compromised, including the healing</p><p>process in the first four to six weeks after surgery, inflammatory changes after trauma,</p><p>recurrent dislocations, and in underlying inflammatory arthropathies. Furthermore, the</p><p>optimal cutoff for the diagnosis of infection is subject to debate. Generally, at this stage of</p><p>the diagnostic algorithm, a test with a high sensitivity is preferred, therefore lower leuko-</p><p>cyte count (e.g. 2000/μL) is preferred, risking overdiagnosis rather than missing an infec-</p><p>tion. The performance of novel biomarkers such as synovial fluid alpha defensin, D‐lactate,</p><p>calprotectin, and leukocyte esterase have been assessed applying different definition crite-</p><p>ria, however leukocyte count showed equal or better performance [31].</p><p>In case of dry joint tap, instillation of saline is not recommended, as the leukocyte count</p><p>analysis is no longer representative due to dilution of synovial fluid. If the synovial fluid</p><p>16 Bone and Joint Infections</p><p>(vascular</p><p>catheter, CIED,</p><p>prosthetic valve)</p><p>, CIED: cardiac</p><p>implantable electronic device</p><p>Figure 2.4. Diagnostic algorithm for periprosthetic joint infection.</p><p>2 Diagnostic Approach in Bone and Joint Infections 17</p><p>analysis is inconclusive or inconsistent with infection, differential diagnoses should be</p><p>considered and further investigated. If the pain remains unexplained, further diagnostic</p><p>procedures depend on the level of suffering of the patient and the stability of the prosthe-</p><p>sis. In case of severe pain with significant compromise of life quality or in case of a loosened</p><p>prosthesis, revision surgery (preferably one‐stage revision) with meticulous intraoperative</p><p>diagnostics is advocated. Biopsy through arthroscopy or mini‐open arthrotomy, depend-</p><p>ing on the affected joint and the expertise of the treating orthopedic surgeon, is an alternative</p><p>in case of a stable prosthesis. However, the diagnostic yield is rarely significant, as the</p><p>representative area (i.e. interface between implant and bone) is not reached. In selected</p><p>cases,</p>