Rates and Predictors of Treatment Failure in Staphylococcus aureus Prosthetic Joint Infections According to Different Management Strategies A Multinational Cohort Study—The ARTHR-IS Study Group

ORIGINAL RESEARCH Ra es and P edic o s o T ea men Failu e in S aphylococcus au eus P os he ic Join In ec ions Acco ding o Di e en Managemen S a egies: A Mul ina ional Coho S udy—The ARTHR-IS S udy G oup Reinaldo Espı ´ndola .Venanzio Vella .Na i idad Beni o .Isabel Mu .Sa a Tedeschi . Eleono a Zampa ini .Johannes G. E. Hend iks .Luisa So lı ´.Osca Mu illo .Lau a Solde ila . Ma hew Sca bo ough .Clai e Sca bo ough .Jan Kluy mans .Ma eo Ca lo Fe a i . Ma hias W. Ple z .Iain Mcnama a .Rosa Escude o-Sanchez .Ced ic A ieux . Cecile Ba aille .F e ´de ´ ic-An oine Dauchy .Wai-Yan Liu .Jaime Lo a-Tamayo .Julia P aena . And ew Us ianowski .Elisa Cinconze .Michele Pelleg ini .Fabio Bagnoli .Jesu ´s Rod ı ´guez-Ban ˜o. Ma ia Dolo es del To o Recei ed: June 23, 2022 / Accep ed: Sep embe 20, 2022 / Published online: Oc obe 15, 2022 ÓThe Au ho (s) 2022 ABSTRACT In oduc ion: Guidelines ha e imp o ed he managemen o p os he ic join in ec ions (PJI). Howe e , i is necessa y o eassess he inci- dence and isk ac o s o ea men ailu e (TF) o S aphylococcus au eus PJI (SA-PJI) including unc ional loss, which has so a been neglec ed as an ou come. Me hods: A e ospec i e coho s udy o SA-PJI was pe o med in 19 Eu opean hospi als be ween 2014 and 2016. The ou come a iable was TF, including ela ed mo ali y, clinical ailu e and unc ional loss bo h a e he ini ial su gical p ocedu e and a e all p ocedu es a 18 mon hs. P edic o s o TF we e iden i ied by logis ic eg ession. Landma k analysis was used Jesu ´s Rod ı ´guez-Ban ˜o and Ma ia Dolo es del To o con ibu ed equally as senio au ho s o he s udy. Membe s o The ARTHR-IS G oup a e lis ed in Acknowledgemen s sec ion. Supplemen a y In o ma ion The online e sion con ains supplemen a y ma e ial a ailable a h ps:// doi.o g/10.1007/s40121-022-00701-0. R. Espı ´ndola In ec ious Diseases and Mic obiology Clinical Uni , Hospi al Uni e si a io Vi gen Maca ena, Se ille, Spain V. Vella E. Cinconze M. Pelleg ini F. Bagnoli GlaxoSmi hKline (GSK), Siena, I aly N. Beni o I. Mu In ec ious Diseases Uni , Hospi al de la San a C eu i San Pau, Ba celona, Spain N. Beni o I. Mu San Pau Ins i u e o Biomedical Resea ch, Ba celona, Spain N. Beni o I. Mu Depa men o Medicine, Uni e si a Au o `noma de Ba celona, Ba celona, Spain S. Tedeschi Depa men o Medical and Su gical Sciences, Uni e si y o Bologna, Bologna, I aly S. Tedeschi E. Zampa ini In ec ious Diseases Uni , IRCCS Azienda Ospedalie o Uni e si a ia di Bologna, Bologna, I aly J. G. E. Hend iks Depa men o O hopaedic Su ge y and T auma, Ma ´xima MC, Eindho en, The Ne he lands .The ARTHR-IS G oup In ec Dis The (2022) 11:2177–2203 h ps://doi.o g/10.1007/s40121-022-00701-0 o a oid immo al ime bias wi h i ampicin when deb idemen , an ibio ics and implan e en ion (DAIR) was pe o med. Resul s: One hund ed wen y cases o SA-PJI we e included. TF a es a e he i s and all su gical p ocedu es pe o med we e 32.8% and 24.2%, espec i ely. A e all p ocedu es, unc- ional loss was 6.0% o DAIR and 17.2% o p os hesis emo al. Va iables independen ly associa ed wi h TF o he i s p ocedu e we e Cha lson C2, haemoglobin 10 g/dL, bac e - aemia, polymic obial in ec ion and addi ional deb idemen (s). Fo DAIR, TF was also associ- a ed wi h a body mass index (BMI) [30 kg/m 2 and delay o DAIR, while i ampicin use was p o ec i e. Fo all p ocedu es, he a iables associa ed wi h TF we e haemoglobin 10 g/ dL, hip ac u e and addi ional join su ge y no ela ed o pe sis en in ec ion. Conclusions: TF emains common in SA-PJI. Func ional loss accoun ed o a subs an ial p opo ion o ea men ailu es, pa icula ly a e p os hesis emo al. Use o i ampicin a e DAIR was associa ed wi h a p o ec i e e ec . Among he isk ac o s iden i ied, anaemia and obesi y ha e no equen ly been epo ed in p e ious s udies. T ial egis a ion: This s udy is egis e ed a clinical ials.go , egis a ion no. NCT03826108. L. So lı ´ Depa men o In ec ious Diseases, Hospi al del Ma , Ins i u Hospi al del Ma d’In es igacions Me `diques (IMIM), Ba celona, Spain L. So lı ´ Uni e si a Pompeu Fab a, Ba celona, Spain O. Mu illo L. Solde ila Depa men o In ec ious Diseases, Hospi al Uni e si a i Bell i ge, IDIBELL, Ba celona, Spain M. Sca bo ough C. Sca bo ough Bone In ec ion Uni , Nu ield O hopaedic Cen e, Ox o d Uni e si y Hospi als NHS Founda ion T us , Ox o d, UK J. Kluy mans Depa men o In ec ion Con ol, Amphia Hospi al, B eda, The Ne he lands J. Kluy mans Depa men o Medical Mic obiology, Uni e si y Medical Cen e U ech , U ech Uni e si y, U ech , The Ne he lands M. C. Fe a i P os he ic-Join Replacemen Uni , Humani as Resea ch Hospi al, Milan, I aly M. W. Ple z Cen e o In ec ious Diseases and In ec ion Con ol, Jena Uni e si y Hospi al, Jena, Ge many I. Mcnama a Depa men o O hopaedics, No olk and No wich Uni e si y Hospi al, No wich, UK R. Escude o-Sanchez Depa men o In ec ious Diseases, Hospi al Uni e si a io Ramo ´n y Cajal, Mad id, Spain R. Escude o-Sanchez J. Rod ı ´guez-Ban ˜o M. D. del To o (&) Cen o de In es igacio ´n en Red de En e medades In ecciosas (CIBERINFEC), Ins i u o de Salud Ca los III, Mad id, Spain e-mail: [email p o ec ed] C. A ieux Depa men o In ec ious Diseases, Cen e Hospi alie Uni e si ai e de Rennes, Rennes, F ance C. Ba aille O hopedic Su ge y Depa men , C oix Rousse Hospi al, Lyon, F ance F.-A. Dauchy Depa men o In ec ious and T opical Diseases, Cen e Hospi alie Uni e si ai e de Bo deaux, Bo deaux, F ance W.-Y. Liu Depa men o O hopaedic Su ge y and T auma, Ca ha ina Hospi al, Eindho en, The Ne he lands 2178 In ec Dis The (2022) 11:2177–2203 G aphical abs ac : W.-Y. Liu Depa men o O hopaedic Su ge y and T auma, Ma ´xima MC, Eindho en, The Ne he lands J. Lo a-Tamayo Depa men o In e nal Medicine, Hospi al Uni e si a io, 12 de Oc ub e, Mad id, Spain J. P aena Clinical Uni o In ec ious Diseases and Mic obiology, Hospi al Uni e si a io Vi gen del Rocı ´o, Se ille, Spain A. Us ianowski In ec ious Diseases Uni , No h Manches e Gene al Hospi al, Manches e , UK J. Rod ı ´guez-Ban ˜oM. D. del To o In ec ious Diseases and Mic obiology Clinical Uni , Hospi al Uni e si a io Vi gen Maca ena, A da D Fed iani 3, 41009 Se ille, Spain In ec Dis The (2022) 11:2177–2203 2179 PLAIN LANGUAGE SUMMARY S aphylococcus au eus is one o he mos i ulen bac e ia and equen ly causes p os he ic join in ec ions. Knowledge o he ea men o his ype o in ec ion has ad anced in ecen yea s, and ea men guidelines ha e led o imp o ed managemen . Typically, he success ul ea - men o hese in ec ions has been de e mined by clinical cu e, ha is, he symp oms o in ec ion ha e disappea ed, bu has no aken in o accoun loss o unc ion (such as signi ican di icul ies walking), which is c i ical o he pa ien ’s qual- i y o li e. Ou aim in his s udy was o e alua e he success o cu en managemen s a egies o S. au eus p os he ic join in ec ion, including eco e y o unc ionali y, and he ac o s ha p edic why some o hese in ec ions a e no cu ed, o iden i y a eas o imp o emen . In a mul ina ional coho o 128 pa ien s wi h S. au eus p os he ic join in ec ion, a es o ea men ailu e we e ound o be high, wi h signi ican a es o loss o unc ion, especially when he p os hesis needed o be emo ed. Loss o unc ion was less equen when he in ec ion was ini ially ea ed wi h su gical cleaning wi h- ou emo al o he p os hesis, e en when his p ocedu e ailed a i s . We ound ha anaemia and obesi y we e associa ed wi h lowe ea men success, and ha he p obabili y o ea men success inc eased when su gical cleaning wi h- ou p os hesis emo al was pe o med ea ly, and when he an ibio ic i ampicin was used in combina ion wi h ano he an ibio ic. Keywo ds: P os he ic join in ec ion; S aphylococcus au eus; Ou come; Clinical ailu e; Func ional ailu e Key Summa y Poin s Why ca y ou his s udy? S aphylococcus au eus has been associa ed wi h a highe a e o p os he ic join in ec ion (PJI) and ea men ailu e compa ed wi h o he pa hogens. Changes in he managemen o S. au eus PJI (SA-PJI) in ecen decades make i necessa y o eassess he incidence and isk ac o s o ea men ailu e, including unc ional loss which has p e iously been neglec ed as an ou come. Wha was lea ned om he s udy? A conside able p opo ion o SA-PJIs ailed a e ini ial su gical ea men s, al hough a subs an ial pa o hem can be escued by addi ional p ocedu es. Signi ican unc ional loss mus be conside ed in addi ion o clinical ailu e, mainly in pa ien s in whom he p os hesis was emo ed, wi h no possibili y o eco e y. In pa ien s managed wi h deb idemen , an ibio ic and p os hesis e en ion (DAIR), he loss o unc ion was lowe e en i his p ocedu e ails. Anaemia and obesi y we e isk ac o s o ea men ailu e ha a e a ely epo ed. The impo ance o some isk ac o s o ea men ailu e was ein o ced, including he p o ec i e ole o i ampicin-based ea men in DAIR. DIGITAL FEATURES This a icle is published wi h digi al ea u es, including [lis digi al ea u es a ailable e.g. a ideo abs ac and slide deck], o acili a e unde s anding o he a icle. To iew digi al J. Rod ı ´guez-Ban ˜oM. D. del To o Depa men o Medicine, School o Medicine, Uni e si y o Se illa, Se ille, Spain J. Rod ı ´guez-Ban ˜oM. D. del To o Biomedicine Ins i u e o Se illa (IBiS), Se ille, Spain P esen Add ess: R. Espı ´ndola Hospi al Uni e si a io de Valme, Se ille, Spain 2180 In ec Dis The (2022) 11:2177–2203 ea u es o his a icle go o h ps://doi.o g/10. 6084/m9. igsha e.21163246. INTRODUCTION S aphylococcus au eus is widely ecognized as a key mic oo ganism causing p os he ic join in ec ion (PJI), and has been associa ed wi h highe a es o ea men ailu e (TF) compa ed wi h o he ae iologies [1]. Howe e , s udies e alua ing TF ha e adi ionally ocused on ou comes ela ed p ima ily wi h con ol o he in ec ion [2–8], while unc ional ou comes ha e been gi en li le conside a ion. A de ini ion o TF ha in eg a es clinical and unc ional aspec s could p o ide a mo e ealis ic measu e- men o he consequences o S. au eus-PJI (SA- PJI). TF a es may a y acco ding o pa ien cha ac e is ics, in ec ion ype (ea ly, delayed o la e) and su gical managemen . Implemen a- ion o he ecommenda ions included in he guidelines o he managemen o PJI [9,10] may ha e posi i ely in luenced changes in he a es and p edic o s o TF in gene al, and pa - icula ly in pa ien s managed wi h deb ide- men , an ibio ics and implan e en ion (DAIR). Fo SA-PJI, howe e , he e is li le published da a on he ou comes associa ed wi h di e en he apeu ic s a egies and hei p edic o s, and he iming and ole o i ampicin in combina- ion wi h o he an ibio ics, pa icula ly in pa ien s unde going DAIR, emain con o e sial. Assessmen s o TF end o be analysed o he i s su gical p ocedu e pe o med o ea he in ec ion, which is impo an o e alua e he ou comes associa ed o he decision abou which p ocedu e mus be pe o med. Howe e , addi ional p ocedu es a e usually pe o med i ha i s p ocedu e ails o cu e he in ec ion, which may escue some pa ien s bu could also a ec he unc ional ou come. E alua ions o TF conside ing all he p ocedu es pe o med a e no usually made. Such an e alua ion would p o ide a mo e global iew o he inal conse- quences o SA-PJI. The objec i es o his s udy we e o p o ide upda ed a es o TF in SA-PJI, aking in o accoun clinical and unc ional aspec s, and o in es iga e he p edic o s o TF acco ding o di e en managemen s a egies. The assess- men o bo h objec i es will be ca ied ou o bo h he i s su gical p ocedu e and all addi- ional p ocedu es pe o med on pa ien s in whom he i s one ailed. METHODS S udy Design, Si es and Pe iod This s udy is pa o he ARTHR-IS p ojec ( eg- is e ed a clinical ials.go : NCT03826108) and was designed o e alua e he incidence, isk ac o s and p edic o s o SA-PJI TF a e p ima y hip and knee a h oplas y. A e ospec i e coho s udy was conduc ed in 19 Eu opean hospi als o iden i y pa ien s olde han 18 yea s who ecei ed a p ima y a h oplas y be ween 1 Janua y 2014 and 31 Decembe 2016 and de eloped pos -su gical knee o hip PJI due o S. au eus wi hin he i s yea a e he p oce- du e. The pa icipa ing si es, loca ed in Spain, I aly, F ance, Ge many, UK and he Ne he - lands, we e selec ed using he CLIN-NET ne - wo k (h ps://www.combac e.com/abou /clin- ne /), based on hei esea ch expe ience and da a collec ion capabili y. The STROBE ecommenda ions we e ol- lowed o he epo ing o he s udy (Supple- men a y Table S1). Pa icipan s Pa ien s wi h pos -su gical hip o knee PJI due o S. au eus diagnosed wi hin he i s yea a e p ima y a h oplas y we e included. The c i e ia used o de ine SA-PJI we e as ollows: p esence o a leas one sign o symp om o PJI, including join pain and/o swelling, o a sinus ac communica ing wi h he p os hesis; and he isola ion o S. au eus om (a) Cone join aspi- a e cul u e, (b) C wo pe ip os he ic issue samples and (c) blood cul u es wi h no o he ob ious sou ce o in ec ion. The pa ien s we e iden i ied by e iewing medical eco ds om In ec Dis The (2022) 11:2177–2203 2181 mic obiological labo a o y da abases, local PJI su eillance da abases and discha ge epo s. Va iables and De ini ions The p ima y ou come a iable was TF un il mon h 18 a e he i s su gical p ocedu e pe - o med, and was analysed sepa a ely bo h o he i s su gical p ocedu e pe o med (mim- icking an in en ion- o- ea analysis o ha p ocedu e) and o all p ocedu es pe o med (including hose pe o med a e ailu e o he i s one). TF was de ined as a composi e a i- able including SA-PJI- ela ed mo ali y, clinical ailu e and unc ional ailu e. Clinical ailu e was de ined as pe sis ence o ecu ence o signs o symp oms o in ec ion. Fo he analysis o he i s p ocedu e, his also included he need o an addi ional cou se o an ibio ics beyond he ini ial one, he need o use long- e m sup- p essi e an ibio ic he apy and need o p os- hesis emo al i no pe o med as he ini ial su gical p ocedu e. Func ional ailu e was de ined as impeded o signi ican ly impai ed walking due o p os he ic loosening o he need o pe o m a Gi dles one p ocedu e o a h odesis. Finally, TF was also analysed in he subg oup o pa ien s who unde wen DAIR as he i s p ocedu e. Po en ial p edic o s o TF we e selec ed acco ding o p e ious s udies [2–8] and addi- ional hypo heses de eloped by he p ojec eam, and a e included in Tables 1and 2. DAIR as p ima y ea men p ocedu e was conside ed app op ia e i i was pe - o med 21 days om onse o PJI symp oms, he e was no sinus ac communica ing wi h he join p os hesis and eplacemen o poly- e hylene o mobile componen s was pe o med acco ding o IDSA guidelines [9]. The de ini- ions o o he a iables a e included in he Supplemen a y Table S2. Da a Collec ion and E hical Aspec s Da a collec ion was supe ised locally by s a wi h ele an expe ise in he ield. Da a we e en e ed in o an anonymized elec onic case epo o m and checked o missing alues and inconsis encies. The s udy was app o ed by he E hics Commi ees a each si e (Supplemen a y Table S5). The need o ob ain w i en in o med consen was wai ed owing o he e ospec i e na u e o he s udy and anonymized da a, excep in he case o he F ench hospi als whe e a le e o non-opposi ion was sen o eligible pa ien s. All pa ien s included in hese cen es he e o e ga e hei au ho iza ion o pa icipa e. S a is ical Analysis Fo bi a ia e analysis o he associa ion o exposu e a iables wi h TF, ela i e isks wi h 95% con idence in e al (CI) we e calcula ed; p alues we e calcula ed by Chi-squa e o Fish- e ’s exac es , as app op ia e. Con inuous a i- ables we e ca ego ized a e analysing o s a i ied associa ions wi h TF. Mul i a iable analyses we e pe o med by logis ic eg ession: he e ec o s udy si e was con olled o using a gene alized linea mixed model in which s udy si es we e conside ed andom e ec s. Va iables wi h p alue 0.15 in bi a ia e analysis, and hose conside ed as po en ially ele an om clinical judgemen , we e en e ed in o he models and selec ed using a manual s epwise backwa d p ocedu e. Va iables wi h p alue 0.1 we e kep in he models. Collinea i y and modi ica ion e ec s be ween a iables we e s udied when clinically sound. The p edic i e abili y o each model was examined by calcula ing hei a eas unde he ecei e ope a ing cha ac e is ic (AUROC) cu es wi h 95% CIs. Fo he e ec o i ampicin in he subg oup o pa ien s who unde wen DAIR as i s su gical ea men , a p opensi y sco e (PS) was calcula ed using a non-pa simo- nious mul i a ia e logis ic eg ession model, in which he ou come a iable was combina ion he apy wi h i ampicin. As u he sensi i i y analyses o he impac o i ampicin he apy, pa simonious mul i a ia e logis ic eg ession models we e pe o med, in which one o wo o he a iables we e emo ed. In addi ion, o a oid immo al ime bias, landma k analysis was used, excluding pa ien s who died o ailed ea men in he i s 21 days a e deb idemen . 2182 In ec Dis The (2022) 11:2177–2203 Table 1 Cha ac e is ics o 128 pa ien s wi h S aphylococ- cus au eus p os he ic join in ec ions (SA-PJI) Va iables No. o pa ien s (pe cen age), excep whe e specified Spain 52 (40.6) F ance 15 (11.7) Ge many 7 (5.5) Uni ed Kingdom 16 (12.5) I aly 17 (13.3) The Ne he lands 21 (16.4) Male sex 65 (50.8) Bo n ab oad 2 (1.5) Age in yea s; median (IQR) 73 (59.3–80.8) Body mass index uni s; median (IQR) 31.2 (25.6–35.1) ASA 3-4 assessmen o p ima y a h oplas y 64 (50%) Cha lson como bidi y index; median (IQR) 1 (1–2) Como bidi ies Ch onic hea ailu e 17 (13.3) Ch onic pulmona y disease 30 (23.4) Diabe es melli us 29 (22.7) Ch onic enal insu ficiency 5 (3.9) Reason o a h oplas y Os eoa h i is 83 (64.8) Hip ac u e 34 (26.6) Os eonec osis 5 (3.9) O he s 6 (4.6) Type o a h oplas y To al hip a h oplas y 50 (39.1) Pa ial hip a h oplas y 27 (21.1) To al knee a h oplas y 47 (36.7) Pa ial knee a h oplas y 4 (3.1) Table 1 con inued Va iables No. o pa ien s (pe cen age), excep whe e specified Me hicillin- esis an S aphylococcus au eus 28 (21.9) Polymic obial in ec ion 36 (28.1) Bac e aemia 25 (19.5) Days om a h oplas y o onse o SA-PJI symp oms; median (IQR) 24 (15–36) Symp oms and signs o SA-PJI Fe e 32 (25.0) Join pain 70 (54.7) Suppu a ion 89 (69.5) Celluli is 38 (29.7) Wound dehiscence 56 (43.8) A icula swelling 18 (14.1) Sinus ac 8 (6.3) Labo a o y da a a diagnosis o SA-PJI; median (IQR) Haemoglobin (g/dL) 10.5 (9.5–11.4) Blood leucocy es (cells/lL) 9600 (7550–12,950) C- eac i e p o ein (mg/L) 84.5 (23.2–224.2) E y h ocy e sedimen a ion a e (mm/h) 68.5 (39.8–92.3) Type o ini ial su gical p ocedu e pe o med o ea SA-PJI Deb idemen and p os hesis e en ion 99 (77.3) Pa ial emo al and eimplan a ion 6 (4.7) One-s age eplacemen and eimplan a ion 4 (3.1) Two-s age eplacemen and eimplan a ion 13 (10.2) In ec Dis The (2022) 11:2177–2203 2183 RESULTS Pa ien Cha ac e is ics and T ea men Failu e Ra es A o al o 130 cases o SA-PJI we e de ec ed, and 128 we e included ( ele an ollow-up da a we e missing o he o he wo). The median numbe o cases pe hospi al was 7 (in e qua ile ange [IQR] 5–9). The median age o pa ien s was 73 yea s (IQR 59–81 yea s); 65 (50.8%) we e males; 77 (60.2%) had hip a h oplas y (50 o al and 27 pa ial) and 51 (39.8%) knee a h o- plas y (47 o al and 4 pa ial). Pa ien cha ac e is ics a e p esen ed in Table 1. In ec- ion- ela ed symp oms s a ed a median o 24 (IQR 15–36) days a e he p ima y a h oplas y (Supplemen a y Fig. S1), while he i s su gical p ocedu e o ea men o in ec ion was pe - o med a median o 4 days (IQR 1–11) a e symp om onse . Bac e aemia occu ed in 25 cases (19.5%). O e all, ou o 128 PJI cases, 28 (21.9%) we e due o me hicillin- esis an S. au- eus (MRSA) s ains. Figu e 1shows pa ien ou comes acco ding o he i s and addi ional su gical p ocedu es pe o med. The a e o TF a e he i s p oce- du e was 32.8% (42 pa ien s; 95% CI 25.2–41.3%). TF was due o clinical ailu e in 27 cases (21.1%), ela ed dea hs in 9 (7%) and loss o unc ion in 6 (4.7%). Median days un il ail- u e was 126 (IQR 34–335). Dea hs occu ed a a median o 21 (IQR 13–48) days a e he i s su gical p ocedu e pe o med. A e u he su gical in e en ions, 11 pa ien s who ailed he i s p ocedu e (9 wi h pe sis en in ec ion, 2 wi h unc ional ailu e due o p os hesis loosening a e DAIR) we e escued. A e 18 mon hs o ollow-up, TF was 24.2% (95% CI 17.5–32.3%). The easons o ailu e we e 9 ela ed dea hs (7.0%), 11 cases o clinical ailu e and 11 cases o unc ional loss (8.5%, espec i ely). Excluding unc ional loss, he ailu e a e was 15.5%. O he 99 pa ien s who ecei ed DAIR as a i s in e en ion o ea he SA-PJI (Fig. 1), 31 (31.3%) ailed ea men due o dea h (n=6), clinical ailu e (n= 23) o loss o unc ion (n= 2). Despi e u he in e en ions, 15 we e s ill ailing a he end o he 18-mon h ollow- up. O he 29 pa ien s who ecei ed o he ypes o i s in e en ions o ea SA-PJI, 11 (37%) ailed ea men , b oken down as ollows: ela- ed dea h (n= 3), clinical ailu e (n= 4) and loss o unc ion (n= 4). No wi hs anding u he in e en ions, 7 we e s ill ailing a he end o he 18-mon hs ollow-up. A summa y o he a es and easons o ea men ailu e is p o ided in Supplemen a y Table S3, including he a e o TF o p os hesis emo al as i s p ocedu e. Table 1 con inued Va iables No. o pa ien s (pe cen age), excep whe e specified Gi dles one p ocedu e (hip esec ion a h oplas y) 6 (4.7) Days om onse o SA-PJI symp oms o fi s su gical p ocedu e pe o med; median (IQR) 4 (1–11) Days o an ibio ic ea men o ea PJI; median (IQR) Empi ical in a enous 2 (1–4) Ta ge in a enous 16 (11–34) Ta ge o al 50 (33–80) To al 73 (56–96) Ta ge o al an ibio ics Ri ampicin 103 (80.5) Fluo oquinolones 73 (63.5) Clindamycin 17 (14.8) T ime hop im–sul ame hoxazole 15 (13.0) Te acyclines 3 (2.6) Linezolid 4 (3.5) O he s 3 (2.6) 2184 In ec Dis The (2022) 11:2177–2203 Table 2 Bi a ia e analysis o po en ial p edic o s o ea men ailu e among pa ien s wi h SA-PJI: a e he fi s su gical p ocedu e, a e DAIR, and a e all su gical p ocedu es pe o med Failu e a e he ini ial su gical p ocedu e (n= 128; ailu es = 42) Failu e a e DAIR as fi s su gical p ocedu e (n= 99; ailu es = 31) Failu e a 18 mon hs a e all p ocedu es pe o med (n= 128; ailu es = 31) Va iables No. ailu e (%) Rela i e isk (95% CI) p- Value No. ailu e (%) Rela i e isk (95% CI) p- Value No. ailu e (%) Rela i e isk (95% CI) p- Value Age C80 yea s 13 (40.6) 1.6 (0.7–3.6) 0.277 8 (38.1) 1.5 (0.5–4.1) 0.450 12 (37.5) 2.4 (1.1–5.8) 0.043 80 yea s 29 (30.2) 23 (29.5) 19 (19.8) Sex Male 22 (33.8) 1.1 (0.5–2.3) 0.800 15 (30.6) 1.07 (0.4–2.5) 0.880 17 (26.2) 1.2 (0.5–2.8) 0.378 Female 20 (31.7) 16 (32.0) 14 (22.2) Cha lson index C2 Yes 21 (47.7) 2.7 (1.3–5.9) 0.009 16 (44.4) 2.6 (1.1–6.1) 0.033 15 (34.1) 2.2 (0.9–5.0) 0.059 No 21 (25.0) 15 (23.8) 16 (19.0) Haemoglobin 10 mg/dl Yes 23 (48.9) 3.1 (1.4–6.7) 0.003 18 (50.0) 3.8 (1.6–9.4) 0.002 18 (38.3) 3.2 (1.4–7.5) 0.005 No 19 (23.5) 13 (20.6) 13 (16.0) Leukocy es C7500/lL Yes 29 (31.9) 0.9 (0.4–2.2) 0.882 19 (27.9) 1.9 (0.5–7.6) 0.383 24 (26.4) 2.3 (0.7–7.4) 0.211 No 10 (33.3) 9 (37.5) 4 (13.3) C- eac i e p o ein C150 mg/L a Yes 18 (36.7) 1.3 (0.6–2.8) 0.478 10 (31.3) 1.0 (0.4–2.5) 0.992 15 (30.6) 1.7 (0.7–3.8) 0.221 No 22 (30.6) 19 (31.1) 15 (20.8) In ec Dis The (2022) 11:2177–2203 2185 pa ien s who died o ailed in he i s 21 days a e deb idemen , use o i ampicin emained p o ec i e o TF [adjus ed ORs (95%CI), 0.22 (0.06–0.80); AUROC cu e o he model, 0.82 (95% CI 0.72–0.91)]. Table 2also p esen s he bi a ia e analysis o isk ac o s o pa ien s who ailed a e all p o- cedu es pe o med. In addi ion o he a iables iden i ied o he i s p ocedu e, TF was also associa ed wi h age [80 yea s, p os hesis emo al as i s p ocedu e, need o pe o m addi ional join su ge y no due o pe sis en in ec ion and non-use o i ampicin and luo- oquinolones in combina ion. On mul i a ia e analysis, hip ac u e [aOR 4.6 (95% CI 1.6–12.9)], haemoglobin le el 10 g/dL [aOR 2.5 (95% CI 1.0–6.6)] and need o pe o m addi ional join su ge y no due o pe sis en in ec ion [aOR 3.2 (95% IC 1.1–8.9)] we e independen ly associa ed wi h TF a 18 mon hs (Table 3). The AUROC o he model o he obse ed da a was 0.80 (95% CI 0.71–0.90). DISCUSSION In his s udy, we ound ha nea ly a hi d o ini ial su gical p ocedu es esul ed in TF. The TF a e dec eased when u he su gical p ocedu es we e pe o med. Impo an ly, we es ima ed he impac o signi ican unc ional loss. When DAIR was used as he i s p ocedu e, e en hough i is a less agg essi e s a egy, addi ional p ocedu es escued a signi ican p opo ion o ini ial ailu es wi hou inc easing loss o unc- ion. In addi ion, he p edic o s o TF o SA-PJIs we e iden i ied, and he ole o i ampicin in pa ien s unde going DAIR was con i med. A e iew o he li e a u e on SA-PJI s udies ocusing on ea men ou comes highligh ed he di icul ies o compa ing di e en s udy esul s owing o he e ogenei y in s udy design, case de ini ions adop ed, leng h o ollow-up and ypes o analyses used (Table 4). O e all, p e iously epo ed TF a es anged om 0% o 16.6% o p os hesis emo al [4,5,8,11] and 13.6% o 63% o DAIR [2,4–8,12–15]. I should also be no ed ha , o he bes o he au ho s’ knowledge, unc ional ou come was no con- side ed a all in p e ious s udies, despi e i being c i ical o he quali y o li e o pa ien s. Func- ional loss is signi ican ly in luenced by he su gical p ocedu es pe o med and his in o - ma ion is he e o e ele an o he decision- making p ocess. When DAIR was analysed as he ini ial p o- cedu e, he TF a e was 31.3% (29.2% wi hou conside ing unc ional loss), which is sligh ly highe han epo ed in mo e ecen obse a- ions [7,12] bu lowe han in olde publica- ions [2–4,6,13] (Table 4). Howe e , he TF a e dec eased o 21.2% a e addi ional p ocedu es, and o 15% i unc ional loss was no aken in o accoun (which is mo e consis en wi h de ini- ions in p e ious epo s). The lowe TF a es o DAIR epo ed by he la es s udies (and by his one) could be a ibu ed o be e pa ien selec ion o his ea men s a egy and o he in ol emen o mul idisciplina y eams in he managemen o PJI. Indeed, Bouaziz e al. [13] ound an o e all TF a e o 42%, which dec eased o 30% when DAIR was pe o med acco ding o he la es guidelines [9]. Ou da a u he sugges ha app op ia e pa ien selec- ion a ou s mo e posi i e ou comes, and ha an ini ial TF can be escued wi hou signi ican unc ional loss in a conside able numbe o pa ien s. The o e all TF a es, including clinical and unc ional upda es, o p os hesis emo al in ou coho (34.4%) may seem ela i ely high when compa ed wi h o he s udies, bu when only clinical cu e was conside ed: he a e o TF o SA-PJI was 17.2%, which is simila o ha ound in o he se ies [4,5,11]. The high p o- po ion o TF a e p os hesis emo al ela ed o unc ional loss is no ewo hy and ein o ces he impo ance o ea ly diagnosis o SA-PJI o inc ease he likelihood o being ea ed wi h DAIR. Since he decision o pe o m DAIR o emo e he p os hesis as i s p ocedu e is s ongly in luenced by pa ien and in ec ion cha ac e is ics, we did no y o compa e he ou comes o he wo p ocedu es as hey a e no compa able. Ins ead, we ocused ou analysis on iden i ying po en ial p edic o s o TF. Rega d- ing he a iables iden i ied, he Cha lson index is a p edic o o su i al and also o p ognosis o many in ec ions; simila ly, o he s udies ha e 2192 In ec Dis The (2022) 11:2177–2203 Table 4 Summa y o published s udies on he ou come and managemen o PJI including C20 cases ocusing on S. au eus Re e ences Yea s o diagnosis and design Types o in ec ion and numbe o pa ien s Numbe o pa ien s and managemen Cu e o ailu e defini ion Ou come Fac o s associa ed wi h ou come B and e al. (1997) [2] 1980–1991 Re ospec i e coho , unicen ic Pos -su gical knee and hip SA-PJI N= 33 (3 MRSA) DAIR (n= 33) Median 28 days in a enous (IV) an ibio ics (be a- lac ams, 91%; ancomycin, 9%) Failu e: pe sis ence o ecu ence o clinical signs o PJI Failu e: 59% a 1 yea 63% a 2 yea s Mul i a iable (HR): DAIR pe o med [2 days a e symp om onse Salgado e al. (2007) [3] 1998–2004 Re ospec i e coho , unicen ic Hip and knee SA-PJI N= 45 (12 MRSA) DAIR (n= 20) Pa ial/ o al 1 T (n=4) 2T(n= 15) Resec ion a h oplas y (n=6) Median 42 days IV an ibio ics (be a-lac ams o ancomycin, 51% wi h i ampicin). In ou DAIR, o al i ampicin-quinolone Failu e: elapse, ein ec ion, dea h- ela ed, o he signs o clinical ailu e Failu e: 38%, 32% in MSSA, 50% in MRSA Follow-up: median 190 days ( ange 4–2279 daus) Mul i a iable: MRSA, TKA, e en ion o p os hesis Vilchez e al. (2010) [14] 2000–2007 Re ospec i e coho , unicen ic Ea ly ( 2 mon hs, 15 days o symp oms) SA-PJI, o al and pa ial hip and knee p os hesis N= 53 (4 MRSA) DAIR (n= 53) Mean 10.6 days i an ibio ics (cloxacillin i MSSA, ancomycin i MRSA) and 88 days o al (le ofloxacin plus i ampicin) Cu e: Absence o in ec ion symp oms, asep ic loosening ha equi ed exchange p os hesis Failu e: 24.5% Follow-up: 2 yea s Mul i a iable (HR): CRP [22 mg/dl Need o 2nd DAIR P os hesis age [25 days In ec Dis The (2022) 11:2177–2203 2193 Table 4 con inued Re e ences Yea s o diagnosis and design Types o in ec ion and numbe o pa ien s Numbe o pa ien s and managemen Cu e o ailu e defini ion Ou come Fac o s associa ed wi h ou come Joulie e al. (2011) [4] 2001–2006 P ospec i e coho , unicen ic ( e e ence) Any o al and pa ial hip and knee SA-PJI N= 95 e aluable (25% MRSA) DAIR (n= 30), 1 T (n= 15), 2 T (n= 25), esec ion (n= 10) Mean 7 days IV an ibio ics (no specified, adap ed o suscep ibili y) and 122.3 days o al (no specified, adap ed o suscep ibili y, 64% combined wi h i ampicin) Cu e: ESR and/o CRP no mal, non-inflamma o y sca wi h no fis ula, no an ibio ics since discha ge and no ein e en ion needed Cu e: DAIR, 57%; 1 T, 94%; 2 T, 86.2%; Resec ion, 34% Follow-up: minimum o 12 mon hs Mul i a iable o cu e: Monomic obial P os hesis emo al Senne ille e al. (2011) [5] 2000–2006 Re ospec i e coho , unicen ic To al hip and knee SA-PJI N= 98 (17.3% MRSA) DAIR (n= 41) (pe o med i symp oms 4 weeks) 1T(n= 14) 2T(n= 26) Resec ion a h oplas y (n=9) A h odesis (n=8) Mean 7 days IV an ibio ics (no specified, adap ed o suscep ibili y) and 3–6 mon hs o al ( i ampicin, n= 68; i ampicin-fluo quinolone, n= 39). Six DAIR wi h supp essi e ea men Cu e: no local o sys emic signs o in ec ion, no need o ein e en ion o new an ibio ic he apy, no in ec ion- ela ed dea hs Cu e: 78.6%; DAIR, 78.0%; 1 T, 100%; 2 T, 84.6%; Resec ion, 44.4%; A h odesis, 62.5% Follow-up: minimum o 2 yea s Mul i a iable o cu e: ASA sco e B2, use o i ampicin- fluo oquinolone 2194 In ec Dis The (2022) 11:2177–2203 Table 4 con inued Re e ences Yea s o diagnosis and design Types o in ec ion and numbe o pa ien s Numbe o pa ien s and managemen Cu e o ailu e defini ion Ou come Fac o s associa ed wi h ou come Lo a-Tamayo e al. (2013) [6] 2003–2010 Re ospec i e coho , mul icen e Ea ly and haema ogenous hip ( o al and pa ial) and knee SA-PJI N= 345 (23% MRSA) DAIR (n= 345) Median days o an ibio ics: 94 in MRSA and 91 in MSSA. Ri ampicin combina ions in 88%: be a- lac ams (13%) o quinolones (75%, mainly le ofloxacin) in MSSA, and glycopep ides (18%, namely ancomycin), co imoxazole (46%), linezolid (24%) o clindamycin (10%) in MRSA Failu e: in ec ion- ela ed dea h, p os hesis eplacemen , need o u he deb idemen s Failu e: 45% (in ec ion- ela ed dea h, 7%) – Ea ly ailu e (30 days), 29% ; Failu e du ing an ibio ic he apy, 32%; – Failu e a e an ibio ic he apy, 39%; Follow-up: 2 yea s Mul i a iable (HR) o ea ly ailu e: male, heuma oid a h i is, bac e aemia, polymic obial, CRP [100 mg/L Failu e du ing he apy: highe age, immunosupp essi e d ugs, MRSA, sinus ac , abno mal adiog aphy, need C2 deb idemen s, no use o i ampicin Failu e a e he apy: haema ogenous in ec ion, deb idemen delay, need o C2 deb idemen s Go ´mez- Junyen e al. (2021) [11] 2003–2010 Re ospec i e coho , mul icen e Hip and knee SA-PJI N= 249 (ea ly, n= 141; haema ogenous, n= 26; ch onic, n= 82) Implan emo al (161 ini ial he apy, 88 sal age) 1T,n= 17 (6.8%) 2T,n= 188 (75.5%) Hip esec ion, n=44 (17.7%) Failu e: local ailu e and/o all-cause mo ali y wi hin 60 days Failu e: 15.6% Local ailu e: 9.3% Mo ali y: 12.8% Follow-up: median 781 days, in e qua ile ange [IQR] 355–1375 days Mul i a iable o ailu e: C2 como bidi ies In ec Dis The (2022) 11:2177–2203 2195 Table 4 con inued Re e ences Yea s o diagnosis and design Types o in ec ion and numbe o pa ien s Numbe o pa ien s and managemen Cu e o ailu e defini ion Ou come Fac o s associa ed wi h ou come Be z e al. (2015) [15] 1996–2012 Re ospec i e coho , unicen ic Hip ( o al and pa ial) PJI N= 29 (12 MSSA, 17 MRSA, 9 s ep ococci) DAIR (n= 38) All exchange o mobile pa s. An ibio ic adap ed o suscep ibili y, wi h median du a ion o 12 weeks (IV o a median o 14 days). Ri ampicin use: 23 (60.5%) Pe sis ence o ecu ence o PJI Failu e: 18.4% MSSA, 9.5% MRSA, 29.4% S ep ococci, 0% Follow-up: minimum o 2 yea s Uni a ia e analysis: no ac o s associa ed wi h ailu e Bouaziz e al. (2018) [13] 2000–2010 Re ospec i e coho , wo cen es Hip and knee MSSA-PJI N= 85 (ea ly, 33; delayed, 11; la e, 45) DAIR (n= 62) An ibio ic no specified, excep o i ampicin Ri ampicin use: ea ly 20 (62%), delayed 7 (63%), la e 26 (58%) Failu e: need o u he su ge y o con ol PJI (o o ea supe in ec ion), ampu a ion o in ec ion- ela ed dea h Failu e: 42% Ea ly, 10 (30%) Delayed, 4 (36%) La e, 23 (51%) Follow-up: 2.8 ±2.2 yea s Mul i a iable (HR) o ailu e: non-compliance wi h su gical IDSA guidelines (namely DAIR pe o med i du a ion o symp oms 3 weeks o join age 30 days, and s able implan wi hou sinus ac ) Lesens e al. (2018) [12] 2010–2014 Re ospec i e coho , mul icen e Hip, knee, shoulde , elbow SA-PJI N= 137 (ea ly acu e, n= 63; ea ly ch onic, n= 26; la e acu e, n= 35; la e ch onic, n= 13) MRSA 27 (19.7%) DAIR (n= 137) Mean du a ion an ibio ic: 12.6 weeks Ri ampicin use: 85 (65%) Ri ampicin plus fluo quinolone: 63 (47.4%) Supp essi e ea men : 14 (10.2%) Failu e: p os hesis emo al, dea h, addi ional deb idemen o cou se o an ibio ics, clinical and mic obiological signs o in ec ion Failu e: 25% Mo ali y: 8.8% Follow-up: 2 yea s Mul i a iable (HR) o ailu e: longe AB du a ion (p o ec i e), i ampicin egimen (p o ec i e), smoking, ea ly in ec ion (p o ec i e) 2196 In ec Dis The (2022) 11:2177–2203 Table 4 con inued Re e ences Yea s o diagnosis and design Types o in ec ion and numbe o pa ien s Numbe o pa ien s and managemen Cu e o ailu e defini ion Ou come Fac o s associa ed wi h ou come Wou huyzen- Bakke e al. (2018) [7] 2005–2015 Re ospec i e coho , wo cen es Ea ly hip and knee MSSA- PJI ea ed wi h i ampicin plus le ofloxacin (n= 40) o moxifloxacin (n= 19) N=58 DAIR (n= 58) Du a ion an ibio ic he apy abou 90 days. IV cloxacillin o flucloxacillin 7–14 days ollowed i ampicin plus fluo quinolone Failu e: need o e ision su ge y and/o supp essi e an imic obial he apy because o pe sis en in ec ion, dea h- ela ed in ec ion, ein ec ion o elapse wi h S. au eus Cu e: 86.4% (87.5% in le ofloxacin g oup, 84.2% in moxifloxacin g oup) Follow-up: minimum 2 yea s No pe o med Mun ˜oz- Gallego e al. (2020) [8] 2016–2017 P ospec i e coho , mul icen e Hip ( o al and pa ial) and knee SA-PJI N= 85 (19 MRSA) Ea ly ( 90 days): 45 Ch onic ([90 days): 21 Haema ogenous: 19 DAIR (n= 55) P os hesis emo al (n= 30) Du a ion and ype o an ibio ic no specified Failu e: all-cause dea h wi hin 90 days, pe sis en o elapsing signs o s aphylococcal in ec ion, need o sal age he apy excep o ex a deb idemen s in he fi s 30 days, supp essi e an imic obial ea men O e all ailu e: 36.4% (DAIR, 47.2%; p os hesis emo al, 16.6%) Follow-up: a leas 1 yea Uni a ia e analysis: Global ailu e associa ed wi h DAIR DAIR: ailu e associa ed wi h delay o deb idemen and no use o i ampicin In ec Dis The (2022) 11:2177–2203 2197 Table 4 con inued Re e ences Yea s o diagnosis and design Types o in ec ion and numbe o pa ien s Numbe o pa ien s and managemen Cu e o ailu e defini ion Ou come Fac o s associa ed wi h ou come ARTHR-IS coho 2014–2016 Re ospec i e coho , mul icen e Hip ( o al and pa ial) and knee pos su gical SA-PJI N= 128 (28 MRSA) DAIR (n= 99) 1T(n= 10) 2T(n= 13) Resec ion (n=6) Median days o an ibio ics: 16 IV and 50 o al. Ri ampicin combina ions in 103 (80.5%); wi h quinolones in 42 (65.8%) Clinical and unc ional ailu e a e he fi s p ocedu e and a 18 mon hs Clinical ailu e: all-cause dea hs in fi s 2 mon hs, SA- PJI- ela ed dea h, pe sis ence o ecu ence o signs o symp oms o in ec ion; need o addi ional cou se o an ibio ics including supp essi e he apy, emo al o he p os hesis i DAIR. Func ional ailu e: significan ly impai ed walking due o p os hesis loosening, Gi dles one o a h odesis O e all clinical and unc ional ailu e a e fi s su gical p ocedu e: 32.8% (DAIR, 33.3%; p os hesis emo al, 37.1%) Conside ing only clinical ailu e: 28.1% Failu e a 18 mon hs: 24.2% (DAIR: 21.2%; p os hesis emo al, 34.4%) Conside ing only clinical ailu e: 15.5% (DAIR: 15.5%; p os hesis emo al: 17.2%) Follow-up: 18 mon hs Mul i a iable analysis: Failu e a e fi s su gical p ocedu e: Cha lson C2, haemoglobin le el 10 mg/dL, bac e aemia, polymic obial in ec ion and need o pe o m addi ional deb idemen In DAIR: hose abo e and BMI [30, and ime be ween symp om onse and DAIR; i ampicin use was p o ec i e A 18 mon hs: haemoglobin le el 10 mg/dL, hip ac u e and need o addi ional join , su gical p ocedu es no ela ed o in ec ion con ol 1Tone-s age eplacemen , 2T wo-s age eplacemen , CRP C- eac i e p o ein, DAIR deb idemen , an ibio ics, and implan e en ion, ESR e y h ocy e sedimen a ion a e, HR haza d a io, IV in a enous, MRSA me hicillin- esis an S. au eus., MSSA me hicillin-suscep ible S. au eus,PJI p os he ic join in ec ion, SA-PJI p os he ic join in ec ion p oduced by S. au eus,TKA o al knee a h oplas y 2198 In ec Dis The (2022) 11:2177–2203 used he ASA index [5]o C2 como bidi ies o simila easons [11]. Anaemia was p e iously iden i ied as a p edic o in one s udy [16], bu was no assessed a all in mos o he o he s. Anaemia is a po en ial ma ke o nu i ional s a us ha can inc ease issue hypoxia o e en indica e a sys emic in lamma o y p ocess. As in ou s udy, bac e aemia, polymic obial in ec ion and he need o addi ional deb idemen we e also p edic o s o TF in o he s udies [6,14]. Since DAIR is he mos equen ini ial p o- cedu e, we also analysed p edic o s o TF in his subg oup. Apa om he abo e a iables, obe- si y and delay in pe o ming DAIR we e also iden i ied as isk ac o s, while he use o i ampicin had a p o ec i e e ec . Obesi y is a isk ac o o PJI and was also ound o be associa ed wi h TF in hip PJI unde going wo- s age eplacemen [17], bu was no e en con- side ed in mos s udies. Obesi y could be asso- cia ed wi h wound complica ions, addi ional deb idemen , impai ed inna e immune esponse and changes in he pha macokine ics o some an imic obial d ugs [18]. Delayed DAIR ollowing onse o PJI symp oms is a known ac o o TF ega dless o he mic oo ganism in ol ed, bu is pa icula ly pe inen in he case o SA-PJI [2,6,12–14]. Es ablishing a h eshold is complex. The 21-day h eshold o pe o ming DAIR ecommended by he guide- lines [9] was based on one small-scale s udy expe ience [19], and a delay o [2 days in pa ien s wi h SA-PJI ea ed wi h be a-lac ams was associa ed wi h inc eased TF in ano he s udy [2] bu no when a i ampicin- luo o- quinolone combina ion was used [6–14,16]. In ou coho , we ound an inc eased isk o each day o delay, suppo ing he ecommenda ion ha deb idemen should be pe o med as ea ly as possible. Ri ampicin in combina ion wi h o he an ibio ics (mainly luo oquinolones) was epo ed o inc ease cu e a es in a small an- domized ial [19] and in obse a ional s udies [5,6,12,20–22]. Howe e , wo ecen me a- analyses ound con o e sial esul s on he ole o i ampicin: one ound no bene i in s aphy- lococcal in ec ions [23] and he o he only a limi ed impac [24]. The s udies included in he me a-analysis had a conside able isk o selec ion and immo al ime bias. In ou s udy, on he o he hand, we ound ha i ampicin combina ions we e associa ed wi h a p o ec i e e ec , e en a e pe o ming sensi i i y and landma k analyses. Al hough p e ious s udies ha e ound highe TF a es in PJIs caused by MRSA compa ed wi h suscep ible s ains [3,15], ou da a did no demons a e his associa ion, which is in line wi h mo e ecen obse a ions [5,6]. Whe he his is due o he use o an i-MRSA d ugs wi h good bioa ailabili y and an i-bio ilm ac i i y, such as linezolid, would equi e u he s udies. Finally, we also analysed he p edic o s o TF a e all p ocedu es had been pe o med. Apa om o he ac o s, hip ac u e inc eased he isk o TF, p obably e lec ing he ail y o he pa ien s a ec ed. In hese pa ien s, he i s su gical app oach was c ucial, since ini ial TF was ollowed by unc ional ailu e in all cases (da a no shown). In a p e ious mul icen e coho s udy o pa ien s wi h hip PJIs, ac u e was also associa ed wi h clinical ailu e and wo se unc ional p ognosis [25]. The isk o TF also inc eased when addi ional join su ge y no due o pe sis en in ec ion was pe o med; his may ha e been due o issue damage, delayed healing o acili a ion o bac e ial supe in ec ions. This s udy has some limi a ions ha should be conside ed when in e p e ing he esul s. The sample size was oo small o in es iga e p edic o s o o he ini ial su gical p ocedu es and may ha e been insu icien o de ec addi- ional p edic o s o TF; i s e ospec i e design limi ed he a ailable a iables; esidual con- ounding is also possible; we did no collec da a abou e hnici y o he pa ien s; inally, some changes in managemen may ha e occu ed du ing he s udy pe iod. Some s eng hs include ha i is a mul ina ional s udy, he de ini ions o TF including clinical and unc ional aspec s and he good p edic i e abili y o he models de eloped. CONCLUSIONS In conclusion, we obse ed ha a conside able p opo ion o SA-PJIs ailed a e ini ial su gical In ec Dis The (2022) 11:2177–2203 2199 ea men s, al hough a subs an ial pa o hem we e eco e ed wi h u he p ocedu es. Signi - ican unc ional loss should be conside ed alongside clinical ailu e, and he impo ance o ce ain isk ac o s o TF was con i med, including he p o ec i e ole o i ampicin- based ea men in DAIR. ACKNOWLEDGEMENTS We hank all s udy pa icipan s o hei col- labo a ion, wi hou which his manusc ip would ha e been impossible. O he membe s o he ARTHR-IS g oup a e: Nienke Cupe us and Giuseppe Man e ´(Julius Cen e o Heal h Sciences and P ima y Ca e, Uni e si y Medical Cen e U ech , U ech , he Ne he lands) collabo a ed in he selec ion p o- cess o he pa icipa ing cen e s. Ana Isabel Sua ´ ez-Ba enechea and Al a o Pascual-He - nandez (Depa men o mic obiology, Hospi al Uni e si a io Vi gen Maca ena), Alba Ri e a (Depa men o Mic obiology, Hospi al de la San a C eu i San Pau), Xa ie C usi and Ma cos Jo da ´n (Depa men o auma ology, Hospi al de la San a C eu i San Pau), Nicolo `Rossi (Depa men o Medical and Su gical Sciences, Uni e si y o Bologna, Bologna, I aly), Tessa an de Ke kho (Depa men o O hopaedic Su ge y & T auma, Ma ´xima MC, Eindho en, he Ne he lands; Depa men o O hopaedic Su ge y & T auma, Ca ha ina Hospi al, Eind- ho en, he Ne he lands), Juan P. Ho cajada, Joan Go ´mez-Junyen and Albe Alie (Hospi al del Ma , Ins i u Hospi al del Ma d’In es iga- cions Me `diques, Uni e si a Pompeu Fab a, Ba celona, Spain), Mi anda an Rijen and Jan- nie Romme (Amphia Hospi al, B eda, Ne he - lands), Juliane Anke (Jena Uni e si y Hospi al—Jena, Ge many), Celia Whi ehouse and Ad ian Jones (No olk and No wich Uni e si y Hospi al—No wich, UK), Ja ie Cobo and Ja ie Mo eno (Hospi al Uni e si a io Ramo ´n y Cajal—Mad id, Spain), Anne Meheu (Cen e Hospi alie Uni e si ai e de Rennes— Rennes, F ance), Clai e Gledel (O hopedic su - ge y depa men , C oix Rousse Hospi al—Lyon, F ance), Pauline Pe eau (Cen e Hospi alie Uni e si ai e de Bo deaux—Bo deaux, F ance), Remco J.A. an Wensen (Depa men o O hopaedic Su ge y & T auma, Ca ha ina Hospi al, Eindho en, he Ne he land), Gab iella Linde ga d (No h Manches e Gene al Hospi- al—Manches e , UK) collabo a ed in he da a collec ion. Funding. This wo k was suppo ed by he Inno a i e Medicines Ini ia i e Join Unde - aking (g an ag eemen No. 115523), COM- BACTE-NET conso ium (Eu opean Union FP7/ 2007–2013 and GlaxoSmi hKline Biologicals SA, as EFPIA pa ne ). R.E, L.S, O. M, R. E-S, J. L–T, J. P, J. R-B and MD. del T a e membe s o he Spanish Ne wo k o Resea ch in In ec ious Diseases (REIPI), suppo ed by Plan Nacional de I?D?i 2013-2016 and Ins i u o de Salud Ca los III, Subdi eccio ´n Gene al de Redes y Cen os de In es igacio ´n Coope a i a, Minis e- io de Ciencia, Inno acio ´n y Uni e sidades, Spanish Ne wo k o Resea ch in In ec ious Diseases (REIPI RD16/0016/0001; 0002; 0005; 0009; 0011; 0015), co- inanced by Eu opean De elopmen Regional Fund ‘‘A way o achie e Eu ope’’, Ope a i e P og am In elligence G ow h 2014-2020. The s udy sponso is also unding he jou nal’s Rapid Se ice Fee. Role o unding sou ce: The unde s had no in luence on he analysis and decision o published; GlaxoSmi hKline Biologicals SA was p o ided he oppo uni y o e iew a e sion o his manusc ip o ac ual accu acy; au ho s a e solely esponsible o inal con en and in e p e a ion. Au ho ship. All named au ho s mee he In e na ional Commi ee o Medical Jou nal Edi o s (ICMJE) c i e ia o au ho ship o his a icle, ake esponsibili y o he in eg i y o he wo k as a whole, and ha e gi en hei app o al o his e sion o be published. Au ho Con ibu ions. Espindola R, Vella V, Rod ı ´guez-Ban ˜o J, Del To o MD con ibu ed in he concep ion and design o he s udy, analysis and in e p e a ion o da a and d a ing he a icle. Beni o N, Mu I, Tedeschi S, Zampa ini E, Hend iks J, So lı ´L, Mu illo O, Solde ila L, Sca bo ough M, Sca bo ough C, Kluy mans J, 2200 In ec Dis The (2022) 11:2177–2203 Fe a i MC, Ple z M, Mcnama a I, Escude o- Sanchez R, A ieux C, Ba aille C, Dauchy F-A, Liu W-Y, Lo a-Tamayo J, P aena J, Us ianowski J con ibu ed in he acquisi ion o da a and e ising he a icle. Cinconze E, Pelleg ini M and Bagnoli F pa icipa ed in he analysis and in e p e a ion o da a and e ision he a icle. All au ho s ga e hei inal app o al o he e - sion o be submi ed. P io P esen a ion. This s udy was p e i- ously p esen ed as a pos e a he 32nd Eu o- pean Cong ess o Clinical Mic obiology & In ec ious Diseases (ECCMIC), held in Lisbon, Po ugal, Ap il 23–26. Disclosu es. Venanzio Vella, Elisa Cinconze, Michele Pelleg ini and Fabio Bagnoli a e employees o he GSK g oup o companies. Venanzio Vella, Michele Pelleg ini and Fabio Bagnoli hold sha es in he GSK g oup o com- panies. Fabio Bagnoli holds pa en s pending and issued pa en s on S aphylococcus au eus accine o mula ions. Venanzio Vella, Elisa Cinconze, Michele Pelleg ini and Fabio Bagnoli decla e no o he inancial o non- inancial ela ionships and ac i i ies. Reinaldo Espindola, Na i idad Beni o, Isabel Mu , Sa a Tedeschi, Eleono a Zampa ini, Johannes G.E. Hend iks, Luisa So lı ´, Osca Mu illo, Lau a Solde ila, Ma hew Sca bo ough, Clai e Sca bo ough, Jan Kluy mans, Ma eo Ca lo Fe a i, Ma hias W. Ple z, Iain Mcnama a, Rosa Escude o-Sanchez, Ced ic A ieux, Cecile Ba aille , F e ´de ´ ic- An oine Dauchy, Wai-Yan Liu, Jaime Lo a- Tamayo, Julia P aena, And ew Us ianowski, Jesu ´s Rod ı ´guez-Ban ˜o and Ma ia Dolo es Del To o ha e no con lic s o decla e. A p esen , he a ilia ion o Reinaldo Espindola is In ec- ious Diseases and Mic obiology Clinical Uni , Hospi al Uni e si a io Vi gen de Valme, Se ille, Spain. Compliance wi h E hics Guidelines. The s udy was app o ed by he E hics Commi ees a each si e. The need o ob ain w i en in o med consen was wai ed due o he e o- spec i e na u e o he s udy and anonymized da a, excep in he case o he F ench hospi als. The app o al o he e hics commi ee o all he pa icipa ing cen e s, wi h he names and e - e ences, a e e lec ed in he able S5 o he supplemen a y ma e ial. Da a A ailabili y. The da ase s gene a ed and analyzed du ing he cu en s udy a e a ailable om he co esponding au ho on easonable eques . Open Access. This a icle is licensed unde a C ea i e Commons A ibu ion-NonComme - cial 4.0 In e na ional License, which pe mi s any non-comme cial use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons licence, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his a icle a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om he copy igh holde . To iew a copy o his licence, isi h p://c ea i ecommons.o g/licenses/by- nc/4.0/. REFERENCES 1. Tande AJ, Pa el R. P os he ic join in ec ion. Clin Mic obiol Re . 2014;27:302–45. h ps://doi.o g/10. 1128/CMR.00111-13. 2. B and CM, Sis unk WW, Du y MC, e al. S a- phylococcus au eus p os he ic join in ec ion ea- ed wi h deb idemen and p os hesis e en ion. Clin In ec Dis 1997;24:914–919. h ps://pubmed. ncbi.nlm.nih.go /9142792/. Accessed 4 June 2021. h ps://doi.o g/10.1093/CLINIDS/24.5.914. 3. Salgado CD, Dash S, Can ey JR, Ma culescu CE. Highe isk o ailu e o me hicillin- esis an S a- phylococcus au eus p os he ic join in ec ions. Clin O hop Rela ed Res. 2007:48–53. h ps://pubmed. ncbi.nlm.nih.go /17534195/. Accessed 3 June 2021. h ps://doi.o g/10.1097/BLO. 0b013e3181123d4e. In ec Dis The (2022) 11:2177–2203 2201