Clinical Infectious Diseases Major Article How to Handle Concomitant Asymptomatic Prosthetic Joints During an Episode of Hematogenous Periprosthetic Joint Infection, a Multicenter Analysis Downloaded from https://academic.oup.com/cid/article/73/11/e3820/5894433 by Fundacio Biblioteca Josep Laporte user on 31 January 2023 Marjan Wouthuyzen-Bakker,1 Marine Sebillotte,2 Cédric Arvieux,2,3 Marta Fernandez-Sampedro,4 Eric Senneville,5 José Maria Barbero,6 Jaime Lora-Tamayo,7 Craig Aboltins,8,9 Rihard Trebse,10 Mauro José Salles,11 Tobias Siegfried Kramer,12,13 Matteo Ferrari,14,15 Joaquín Garcia-Cañete,16 Natividad Benito,17,18 Vicens Diaz-Brito,19 Maria Dolores del Toro,20 Matthew Scarborough,21 and Alex Soriano22; on behalf of the European Society of Clinical Microbiology and Infectious Diseases (ESCMID) Study Group for Implant-Associated Infections (ESGIAI) 1 Department of Medical Microbiology and Infection Prevention, University of Groningen, University Medical Center Groningen, the Netherlands, 2Department of Infectious Diseases and Intensive Care Medicine, Rennes University Hospital, Rennes, France, 3Great West Reference centers for Complex Bone and Joint Infections (CRIOGO), Rennes, France, 4Service of Infectious Diseases, Hospital Universitario Marqués de Valdecilla, University of Cantabria, Instituto de investigación sanitaria Valdecilla (IDIVAL), Santander, Spain, 5Department of Infectious Diseases, Lille, University Hospital Gustave Dron Hospital, Tourcoing, France, 6Department of Internal Medicine. Hospital Universitario Principe de Asturias, Madrid, Spain, 7Department of Internal Medicine. Hospital Universitario 12 de Octubre. Instituto de Investigación i+12. Madrid, Spain, 8The Department of Infectious Diseases, Northern Health, Melbourne, Australia, 9The University of Melbourne, Northern Clinical School, Melbourne, Australia, 10Service for Bone Infections, Valdoltra Orthopaedic Hospital, the Faculty of Medicine, University of Ljublijana, Ankaran, Slovenia, 11Santa Casa de São Paulo School of Medical Sciences and Musculoskeletal infection group, Federal University of São Paulo, Brasil, 12Institute for hygiene and environmental medicine Charité-Universitätsmedizin Berlin, Germany, 13Evangelisches Waldkrankenhaus Spandau, Berlin Germany, 14Department of Biomedical Sciences, Humanitas University, Milan, Italy, 15Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Humanitas Research Hospital, Milan, Italy, 16Department of Internal Medicine-Emergency, Instituto de Investigación Sanitaria (IIS)-Fundación Jiménez Díaz, Universidad Autónoma de Madrid (UAM), Madrid, Spain, 17Infectious Diseases Unit, Hospital de la Santa Creu i Sant Pau, Institut d’Investigació Biomèdica Sant Pau, Barcelona, Spain, 18Department of Medicine. Universitat Autònoma de Barcelona, Spain, 19Infectious Diseases Unit, Parc Sanitari Sant Joan de Deu, Sant Boi, Barcelona, Spain, 20Unidad Clínica de Enfermedades Infecciosa y Microbiología. Universidad de Sevilla. Instituto de Biomedicina de Sevilla (IBIS), Sevilla, Spain, 21Bone Infection Unit, Nuffield Orthopaedic Centre, Oxford University Hospitals National Health Service (NHS) Foundation Trust, Oxford, England, and 22Service of Infectious Diseases, Hospital Clínic, University of Barcelona, Barcelona, Spain Background. Prosthetic joints are at risk of becoming infected during an episode of bacteremia, especially during Staphylocococcus aureus bacteremia. However, it is unclear how often asymptomatic periprosthetic joint infection (PJI) occurs, and whether additional diagnostics should be considered. Methods. In this multicenter study, we retrospectively analyzed a cohort of patients with a late acute (hematogenous) PJI be- tween 2005–2015 who had concomitant prosthetic joints in situ. Patients without at least 1 year of follow-up were excluded. Results. We included 91 patients with a hematogenous PJI and 108 concomitant prosthetic joints. The incident PJI was most fre- quently caused by Staphylococcus aureus (43%), followed by streptococci (26%) and Gram-negative rods (18%). Of 108 concomitant prosthetic joints, 13 were symptomatic, of which 10 were subsequently diagnosed as a second PJI. Of the 95 asymptomatic prosthetic joints, 1 PJI developed during the follow-up period and was classified as a “missed” PJI at the time of bacteremia with S. aureus (1.1%). Infected prosthetic joints were younger than the noninfected ones in 67% of cases, and prosthetic knees were affected more often than prosthetic hips (78%). Conclusions. During an episode of hematogenous PJI, concomitant asymptomatic prosthetic joints have a very low risk of being infected, and additional diagnostic work-up for these joints is not necessary. Keywords. periprosthetic joint infection; late acute; hematogenous; asymptomatic; bacteremia. A serious complication of total joint replacement is the develop- [4, 5]. According to registry observations, patients with a pros- ment of a periprosthetic joint infection (PJI). This complication thetic joint in situ are at risk for developing a late acute (he- occurs in around 1–2% of patients after primary arthroplasty matogenous) PJI at a rate of approximately 0.07% per life year and around 10% after revision arthroplasty [1–3]. Infection of the prosthesis, with the highest risk observed in knees [6]. mostly occurs in the postsurgical period, but may develop at Moreover, during an episode of bacteremia, especially in the a later stage as well, either as a late chronic or late acute PJI case of Staphylocococcus aureus bacteremia, it has been sug- gested that prosthetic joints have a 20–30% chance of becoming Received 15 June 2020; editorial decision 6 August 2020; published online 18 August 2020. Correspondence: M. Wouthuyzen-Bakker, Infectious Disease Specialist, University Medical secondarily infected [7–12]. It is well established that an addi- Center Groningen, Department of Medical Microbiology and Infection Prevention, Hanzeplein 1, tional diagnostic work-up and early surgical debridement are 9713 GZ, Groningen, The Netherlands (m.wouthuyzen-bakker@umcg.nl). required in symptomatic joints with a high clinical suspicion of Clinical Infectious Diseases®  2021;73(11):e3820–4 © The Author(s) 2020. Published by Oxford University Press for the Infectious Diseases Society PJI, but it is unclear how an asymptomatic joint should be ap- of America. This is an Open Access article distributed under the terms of the Creative Commons proached in this regard. A missed diagnosis of an acute PJI may Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. result in the development of a chronic PJI, treatment of which is https://doi.org/10.1093/cid/ciaa1222 likely to mandate removal of the implant [13]. Timely diagnosis e3820 • cid 2021:73 (1 December) • Wouthuyzen-Bakker et al is thus of extreme importance. Patients who are diagnosed with Paents with a hematogenous PJI N=308 a PJI and who have concomitant prosthetic joints in situ ap- pear to have the highest risk for having a second PJI [14–16]. 1 prosthec joint N=217 Therefore, we retrospectively analyzed a cohort of patients with Paents with ≥ 1 concomitant prosthec joint(s) hematogenous PJI who also had 1 or more concomitant pros- N=91 thetic joints in situ at the time of PJI diagnosis. We evaluated how many asymptomatic joints were diagnosed with PJI at the Number of concomitant prosthec joints time of presentation and how many developed a PJI during the N=108 follow-up period. Downloaded from https://academic.oup.com/cid/article/73/11/e3820/5894433 by Fundacio Biblioteca Josep Laporte user on 31 January 2023 Symptomac Asymptomac N=13 N=95 MATERIAL AND METHODS PJI during follow-up PJI N=4 Study Design and Inclusion Criteria N=10 This study was performed as part of a large, international, Same pathogen, N=2 Other pathogen, N=2 multicenter observational study in which data from con- (1 reinfecon, 1 relapse) secutive patients with hematogenous PJI between January 2005 and December 2015 were collected and retrospectively Figure 1. Flow chart for inclusion. Abbreviation: PJI, periprosthetic joint infection. evaluated. A PJI was defined according to the adapted diag- nostic criteria of the Musculoskeletal Infection Society [17], Characteristics of Hematogenous PJIs requiring at least 2 positive, intraoperative tissue cultures Table 1 shows the baseline and clinical characteristics of the with phenotypically identical microorganisms. A hematog- 91 patients. The hematogenous PJI episode mainly manifested enous PJI was defined by the sudden onset of symptoms and in knees (69%), and the infection was most commonly due to by signs of arthritis occurring more than 3 months after the Staphylococcus aureus (43%), followed by streptococci (26%) index arthroplasty in a previously asymptomatic prosthetic and Gram-negative rods (18%). The portal of entry for the bac- joint. Patients with a sinus tract, culture-negative PJIs, and teremia was identified in 46% of cases, and endocarditis was a follow-up of less than 1 year were excluded. Patients were diagnosed in 5%. Patients with S. aureus PJI received an ech- included in the final analysis if they had at least 1 other con- ocardiogram in 17/39 (44%) cases. At the time of clinical pre- comitant prosthetic joint in situ at the time of clinical pre- sentation, blood cultures were positive in 32/91 cases (35%) sentation. Informed consent was obtained when required by and negative in 29/91 cases (32%), and no blood cultures were the ethics committee of the participating center. Variables obtained in 30/91 cases (33%). Most of the patients (77%) were relating to patient characteristics, clinical presentation, diag- treated with debridement, antibiotics, and implant retention. nostics, microbiology, surgical intervention, antibiotic treat- ment, and outcome were collected and analyzed. Analysis of Concomitant Prosthetic Joints Of the 91 patients with a hematogenous PJI, 108 concomitant Statistical Analysis prosthetic joints were analyzed, including 56 knees, 44 hips, A chi-square test (or a Fisher exact test, when appropriate) 3 shoulders, and 5 other joints. Of these, 13/108 (12%) were was used to analyze the difference between groups for cat- symptomatic at the time of presentation and 95/108 (88%) egorical variables, and a student t-test (or Mann-Witney were asymptomatic (Figure 1). Of the 13 symptomatic joints, U test, when data were not normally distributed) was used 7 patients experienced acute pain of the joint (54%), 2 patients for continuous variables. All analyses were 2-tailed and P experienced acute or chronic pain (15%), and 4 patients expe- values < .05 were considered statistically significant. The rienced chronic pain (31%). All of the 13 symptomatic joints statistical analysis was performed using SPSS, version 23.0 had clinical signs of infection on physical examination, and (SPSS Inc., Chicago, IL). 10 of these 13 patients were diagnosed as having a second PJI (77%). In the other 3 patients, a PJI was ruled out: all of these RESULTS had chronic pain, which was attributed to mechanical reasons. Figure 1 shows the inclusion flowchart of patients with a he- The median time of follow-up of the 95 asymptomatic joints matogenous PJI and the number of concomitant prosthetic was 52 months (range, 13–130 months). There were 4 cases joints. Of 308 consecutive patients with hematogenous PJIs, that developed a PJI during follow-up (4.2%); all were caused 91 (29.5%) had at least 1 additional prosthetic joint in situ, re- by S. aureus. In 2 out of 4 cases, the original PJI was caused by sulting in a total of 108 concomitant prostheses for inclusion Streptococcus pyogenes, and was thus considered as a new hema- in the final analysis (1 joint each in 77 patients, 2 joints each in togenous infection. In the other 2 cases, S. aureus was the causa- 11 patients, and 3 joints each in 3 patients). tive microorganism in the original PJI as well. These 2 cases were Asymptomatic Prostheses During Bacteremia • cid 2021:73 (1 December) • e3821 Table 1. Patient Characteristics in Hematogenous Periprosthetic Joint February 2008 with a hematogenous PJI of the right knee caused Infection by S. aureus. Blood cultures were positive during clinical presen- tation. After a failed surgical debridement, the patient underwent Patient characteristics, n = 91 n/N, % a 2-stage exchange procedure and was treated with antibiotics Baseline characteristics accordingly. The patient presented with an acute PJI of the left Gender, male 37/91, 40.7% Age >80 years 23/91, 25.3% knee prosthesis 4 years later, also due to S. aureus. Although the BMI >30 29/56, 51.8% antibiogram was the same, as the patient had not experienced ASA classification ≥III 35/78, 44.9% any chronic pain during this 4-year interval, the infection was Medical history considered as a new hematogenous PJI and not as a relapse of Downloaded from https://academic.oup.com/cid/article/73/11/e3820/5894433 by Fundacio Biblioteca Josep Laporte user on 31 January 2023 Hypertension 48/90, 53.3% an unrecognized infection in 2008. The second patient was diag- Diabetes mellitus 16/88, 18.2% Ischemic heart disease 14/91, 15.4% nosed in September 2011 with a late acute S. aureus PJI of the Rheumatoid arthritis 13/91, 14.3% right hip. This patient had positive blood cultures at presentation. Heart failure 9/91, 9.9% Endocarditis was excluded by transesophageal echocardiography. COPD 8/91, 8.8% The patient was successfully managed with surgical debridement Chronic renal insufficiency 7/91, 7.7% and a rifampin-based antibiotic regimen, consisting of an intra- Active malignancy 6/91, 6.6% venous beta-lactam antibiotic, during the first 4 weeks, and was Liver cirrhosis 3/91, 3.3% Medication treated after that with fusidic acid as the final codrug of rifampin. Oral anticoagulant 16/91, 17.6% The left hip remained asymptomatic at that time. However, after Immune-suppressive drugs 14/91, 15.4% stopping antibiotic treatment in February 2012, the left hip be- Characteristics infected prosthetic joint came symptomatic and S. aureus was subsequently cultured in Knee 63/91, 69.2% this joint, with an identical antibiogram as the initial infection Hip 26/91, 28.6% (both a penicillin-sensitive strain). This PJI was considered as a Indication prosthesis: Osteoarthritis 71/91, 78% relapse of a previously undiagnosed infection. Rheumatoid arthritis 12/91, 13.2% Altogether, only 1 out of the 95 asymptomatic prosthetic Fracture 4/91, 4.4% joints was classified as an unrecognized PJI at the time of the Revision prosthesis 30/90, 33.3% first clinical presentation (1.1%). When considering only S. au- Cemented stem 45/60, 75% reus PJIs, this percentage was 2.7% (1 out of 37); in those with Clinical presentation Duration of symptoms >10 days 33/85, 38.8% documented S. aureus bacteremia, it was 5.6% (1 out of 18). Temperature >38.5°C 20/85, 23.5% Of the 91 patients with at least 1 asymptomatic joint in CRP >150 mg/L 45/79, 56.9% situ, 11 received additional diagnostics: 1 patient underwent Leucocytes >17 cells/ µL 11/85, 12.9% joint aspiration and 10 patients underwent nuclear imaging Positive blood cultures 32/91, 35.1% (fluuorodeoxyglucose [FDG]-positron emission tomography Negative blood cultures 30/91, 32.9% [PET] computed tomography in 4 cases and white blood cell No blood cultures taken 29/91, 31.8% Endocarditis 3/91, 3.3% scintigraphy in 6 cases). In all 11 cases, there were no signs of Portal of entry bacteremia identified 42/91, 46.1% PJI on imaging and synovial fluid analysis, and none of these Microorganism patients were diagnosed with a PJI during follow-up. Staphylococcus aureus 38/91, 41.8% Methicillin resistant 5/91, 5.5% Which Prosthetic Joint Gets Infected Streptococcus species 23/91, 25.3% We additionally investigated whether the type of joint or the age Gram-negative rods 14/91, 15.4% Enterococcus species 5/91, 5.5% of the prosthesis influenced the risk of infection. In the majority Surgical strategy of cases (67%), the infected prosthetic joint was younger than the DAIR 70/91, 76.9% noninfected joint, with a median prosthesis age of 4.5 (interquar- More than 1 DAIR performed 13/70, 18.6% tile range, 1.5–11.8) versus 6.7 (interquartile range, 2.8–14.3) years, Mobile components exchanged 30/70, 42.8% respectively (P = .04; Figure 2). In 27 patients with both knee and Revision surgery 18/91, 19.8% hip prostheses in situ (n = 27), the knee prosthesis became infected No surgery 3/91, 3.3% Abbreviations: ASA, American Society of Anesthesiologist; BMI, body mass index; COPD, in 21/27 (78%) cases, and the hip prosthesis became infected in chronic obstructive pulmonary disease; CRP, C-reactive protein; DAIR, debridement, anti- 6/27 (22%). In no cases were both joints infected. biotics, and implant retention. DISCUSSION evaluated in more detail to decipher whether these were poten- The management of concomitant asymptomatic implants in tial, unrecognized infections at the time of the first PJI episode. patients with a documented PJI is not widely addressed in lit- The first patient, with a bilateral knee prosthesis, presented in erature. Timely diagnosis of PJI is important, in particular for e3822 • cid 2021:73 (1 December) • Wouthuyzen-Bakker et al of surgery. However, it should be noted that in this particular study, PJIs occurring within the first 3 months after the index surgery were included, making it difficult to determine whether the PJI was the cause or the consequence of the bacteremia. Our study has a few limitations. First, we analyzed only pa- tients with a hematogenous PJI who had at least 1 other pros- thetic joint in situ. However, it is known that patients with a history of PJI have a higher chance of having a second PJI, especially during an episode of bacteremia [14–16]. Thus, Downloaded from https://academic.oup.com/cid/article/73/11/e3820/5894433 by Fundacio Biblioteca Josep Laporte user on 31 January 2023 even in this “high-risk” group, the chance of having a PJI in an asymptomatic joint at the time of having a hematogenous PJI is very low. A further limitation of our study is that not Figure 2. Prosthesis age and occurrence of hematogenous PJI. Abbreviation: PJI, periprosthetic joint infection. all hematogenous PJIs had positive blood cultures at the time of clinical presentation. Although this may call into question the hematogenous origin of infection, all diagnosed cases had hematogenous PJIs, where early surgical debridement without a sudden onset of acute symptoms and signs of arthritis, and removing the implant is commonly recommended. In our anal- were entirely asymptomatic prior to clinical presentation in ysis of 91 patients with a hematogenous PJI and 108 concomi- the hospital. This, together with the microorganisms isolated tant prosthetic implants, we demonstrate a very low incidence (which were mainly virulent ones: S. aureus, ß-hemolytic (1.1%) of a second PJI in prosthetic joints that have no clinical streptococci, and Gram-negative rods), makes a hematoge- signs or symptoms of infection. Although we cannot rule out nous origin highly likely. the possibility that asymptomatic infection is being eradicated In conclusion, in patients with bacteremia or with a hema- by the antibiotic treatment used for the index PJI, our results togenous PJI, all concomitant prosthetic joints should be care- suggest that performing additional diagnostics or surgical inter- fully examined. If symptoms or signs of infection are present, ventions is not indicated in asymptomatic prosthetic joints. appropriate additional diagnostics should be undertaken. If Our results are supported by previous analyses. Tande et al clinical symptoms of infection are absent, the chance of PJI is [7] evaluated 85 patients with a S. aureus bacteremia who had very low and additional diagnostics are unnecessary. a prosthetic joint in situ. A PJI was diagnosed in 39/139 (28%) arthroplasties, and 38 of these had clinical signs of infection. In Note line with our results, of the 100 cases in which a PJI was not diag- Potential conflicts of interest. T. S. K. reports personal fees from Infectopharm, outside the submitted work. R. T. has received consulting nosed, only 4% of patients developed a PJI during follow-up, and fees from Medacta, speaker fees from Zimmer Biomet and Link, and grants all were caused by S. aureus. Since the range of developing a PJI from Ceramtec, and is President of European Bone and Joint Infection during follow-up varied between 174 and 670 days, it is unclear Society, outside the submitted work. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of whether these PJIs represented a relapse of infection or a new Potential Conflicts of Interest. Conflicts that the editors consider relevant to hematogenous infection with the same microorganism, which the content of the manuscript have been disclosed. would be reasonable in persistent nasal carriers of S. aureus [18, 19]. Dufour et al [8] recently described a similar observation. References Out of the 143 evaluated arthroplasties in their study, 19% de- 1. Bozic KJ, Kurtz SM, Lau E, Ong K, Vail TP, Berry DJ. The epidemiology of re- veloped a PJI during an episode of S. aureus bacteremia. All of vision total hip arthroplasty in the United States. J Bone Joint Surg Am 2009; 91:128–33. these patients experienced pain or swelling of the affected joint 2. Bozic KJ, Kurtz SM, Lau E, et al. The epidemiology of revision total knee during their hospital stay. No additional PJIs occurred after a arthroplasty in the United States. Clin Orthop Relat Res 2010; 468:45–51. median follow-up of 261 days in the asymptomatic prosthetic 3. Saleh A, Guirguis A, Klika AK, Johnson L, Higuera CA, Barsoum WK. Unexpected positive intraoperative cultures in aseptic revision arthroplasty. J joints. As with our data, these articles suggest that although the Arthroplasty 2014; 29:2181–6. risk of developing a PJI during S. aureus bacteremia is high, it is 4. Zeller V, Kerroumi Y, Meyssonnier V, et al. Analysis of postoperative and hema- togenous prosthetic joint-infection microbiological patterns in a large cohort. J unlikely that asymptomatic implants are infected and that diag- Infect 2018; 76:328–34. nostic measures on these joints can reasonably be avoided. 5. Manning L, Metcalf S, Clark B, et al. Clinical characteristics, etiology, and in- itial management strategy of newly diagnosed periprosthetic joint infection: a The observation that knees are more likely to become in- multicenter, prospective observational cohort study of 783 patients. Open Forum fected than other joints is replicated in several observational Infect Dis 2020; 7:1–10. 6. Huotari K, Peltola M, Jämsen E. The incidence of late prosthetic joint infections: a studies on hematogenous PJIs [6, 7, 20]. The finding of younger registry-based study of 112 708 primary hip and knee replacements. Acta Orthop implants being more prone to infection was also reported by 2015; 86:321–5. Honkanen et al [12], demonstrating that the risk for devel- 7. Tande AJ, Palraj BR, Osmon DR, et al. Clinical presentation, risk factors, and out- comes of hematogenous prosthetic joint infection in patients with Staphylococcus oping a PJI was highest for bacteremias occurring within a year aureus bacteremia. Am J Med 2016; 129:221.e11–20. Asymptomatic Prostheses During Bacteremia • cid 2021:73 (1 December) • e3823 8. Dufour S, Porioth L, Chirouze C, et al. Staphylococcus aureus bloodstream infec- 15. Zeller V, Dedome D, Lhotellier L, Graff W, Desplaces N, Marmor S. Concomitant tion in patients with prosthetic joints in the prospective VIRSTA cohort study: multiple joint arthroplasty infections: report on 16 cases. J Arthroplasty 2016; frequency and time of occurrence of periprosthetic joint infection. Open Forum 31:2564–8. Inf Dis 2019; 6:1–4. 16. Komnos GA, Manrique J, Goswami K, et al. Periprosthetic joint infection in pa- 9. Lalani T, Chu VH, Grussemeyer CA, et al. Clinical outcomes and costs among tients who have multiple prostheses in place: what should be done with the silent patients with Staphylococcus aureus bacteremia and orthopedic device infections. prosthetic joints. J Bone Joint Surg Am 2020; 102:1160–8. Scand J Infect Dis 2008; 40:973–7. 17. Parvizi J, Gehrke T; International Consensus Group on Periprosthetic Joint 10. Murdoch DR, Roberts SA, Fowler VG Jr, et al. Infection of orthopedic prostheses Infection. Definition of periprosthetic joint infection. J Arthroplasty 2014; after Staphylococcus aureus bacteremia. Clin Infect Dis 2001; 32:647–9. 29:1331. 11. Sendi P, Banderet F, Graber P, Zimmerli W. Periprosthetic joint infection fol- 18. Wertheim HF, Vos MC, Ott A, et al. Risk and outcome of nosocomial lowing Staphylococcus aureus bacteremia. J Infect 2011; 63:17–22. Staphylococcus aureus bacteraemia in nasal carriers versus non-carriers. Lancet 12. Honkanen M, Jämsen E, Karppelin M, Huttunen R, Eskelinen A, Syrjänen J. 2004; 364:703–5. Downloaded from https://academic.oup.com/cid/article/73/11/e3820/5894433 by Fundacio Biblioteca Josep Laporte user on 31 January 2023 Periprosthetic joint infections as a consequence of bacteremia. Open Forum 19. von Eiff C, Becker K, Machka K, Stammer H, Peters G. Nasal carriage as a Infect Dis 2019; 6:1–6. source of Staphylococcus aureus bacteremia. Study group. N Engl J Med 2001; 13. Barberan J, Aguilar L, Carroquino G, et al. Conservative treatment of staph- 344:11–6. ylococcal prosthetic joint infections in elderly patients. Am J Med 2006; 20. Wouthuyzen-Bakker M, Sebillotte M, Lomas J, et al; European Society of Clinical 119:993e7–10. Microbiology and Infectious Diseases Study Group for Implant-Associated 14. Haverstock JP, Somerville LE, Naudie DD, Howard JL. Multiple periprosthetic Infections (ESGIAI). Clinical outcome and risk factors for failure in late acute joint infections: evidence for decreasing prevalence. J Arthroplasty 2016; prosthetic joint infections treated with debridement and implant retention. J 31:2862–6. Infect 2019; 78:40–7. e3824 • cid 2021:73 (1 December) • Wouthuyzen-Bakker et al