Open access Protocol Multicentre, randomised, open-­label, phase IV–III study to evaluate the efficacy of cloxacillin plus fosfomycin versus cloxacillin alone in adult patients with methicillin-­susceptible Staphylococcus aureus bacteraemia: study protocol for the SAFO trial Sara Grillo,1,2 Guillermo Cuervo ‍ ‍,1,2 Jordi Carratala,1,3 Rafael San-­Juan,4,5 Jose M Aguado,4,6 Laura Morata,7,8 Silvia Gomez-­Zorrilla,9,10 Joaquín López-­Contreras,11,12 Oriol Gasch ‍ ‍,13,14 Aina Gomila-­Grange,14,15 Simona Iftimie,16 Graciano Garcia-­Pardo,17 Esther Calbo,18,19 Lucía Boix-­Palop,18,19 Isabel Oriol,20 Alfredo Jover-­Sáenz,21,22 Luis Eduardo López-­Cortés,23,24 Gorane Euba,25,26 Malen Aguirregabiria,26,27 Maria Jose Garcia-­Pais,28,29 Francesca Gioia,30,31 Jose Ramón Paño,32,33 Maria Luisa Pedro-­Botet,34,35 To cite: Grillo S, Cuervo G, Rosa Maria Benítez,34,35 Maria Teresa Pérez-­Rodríguez,36 Yolanda Meije,37,38 Carratala J, et al. Multicentre, Maria Belén Loeches-­Yagüe,39 Gertrudis Horna,2 Damaris Berbel,2,40 randomised, open-­label, phase Maria Ángeles Domínguez,2,40 Ariadna Padullés,2,41 Sara Cobo,2,41 Pilar Hereu,2,42 IV–III study to evaluate the Sebastian Videla,2,42 Cristian Tebe,3,43 Natàlia Pallarés,3,44 Josep M Miro,7,8 efficacy of cloxacillin plus fosfomycin versus cloxacillin Miquel Pujol,1,2 for the SAFO study group and the Spanish Network for Research in alone in adult patients with Infectious Diseases (REIPI) methicillin-­susceptible Staphylococcus aureus bacteraemia: study protocol for the SAFO trial. BMJ Open ABSTRACT 2021;11:e051208. doi:10.1136/ Strengths and limitations of this study bmjopen-2021-051208 Introduction Methicillin-­susceptible Staphylococcus aureus (MSSA) bacteraemia is a frequent condition, ►► Prepublication history and ►► The primary endpoints are strong composite out- with high mortality rates. There is a growing interest supplemental material for this comes that will assess mortality, clinical and in identifying new therapeutic regimens able to reduce paper is available online. To microbiological failure at 7 and 90 days after therapeutic failure and mortality observed with the view these files, please visit randomisation. standard of care of beta-­lactam monotherapy. In vitro the journal online (http://​dx.​doi.​ ►► The multicentre nature of the study supports the org/​10.​1136/​bmjopen-​2021-​ and small-­scale studies have found synergy between generalisability of the results. 051208). cloxacillin and fosfomycin against S. aureus. Our aim ►► A blinded adjudication committee will evaluate the is to test the hypothesis that cloxacillin plus fosfomycin key study endpoints and mitigate the observer bias Received 15 March 2021 achieves higher treatment success than cloxacillin alone in Accepted 16 July 2021 inherent in the open-­label design. patients with MSSA bacteraemia. ►► Given the increased risk of sodium overload, pa- Methods We will perform a superiority, randomised, tients with cardiac failure and hepatic cirrhosis will open-­label, phase IV–III, two-­armed parallel group (1:1) be excluded. © Author(s) (or their clinical trial at 20 Spanish tertiary hospitals. Adults (≥18 employer(s)) 2021. Re-­use years) with isolation of MSSA from at least one blood permitted under CC BY-­NC. No culture ≤72 hours before inclusion with evidence of commercial re-­use. See rights infection, will be randomly allocated to receive either Organ Failure Assessment score, afebrile and with negative and permissions. Published by cloxacillin 2 g/4-­hour intravenous plus fosfomycin 3 g/6-­ blood cultures for MSSA at day 7. (2) Treatment success BMJ. hour intravenous or cloxacillin 2 g/4-­hour intravenous at test of cure (TOC) visit: patient alive and no isolation of For numbered affiliations see alone for 7 days. After the first week, sequential treatment MSSA in blood culture or at another sterile site from day 8 end of article. and total duration of antibiotic therapy will be determined until TOC (12 weeks after randomisation). Correspondence to according to clinical criteria by the attending physician. We assume a rate of treatment success of 74% in the Dr Guillermo Cuervo; Primary endpoints: (1) Treatment success at day 7, a cloxacillin group. Accepting alpha risk of 0.05 and beta ​guillermo.​cuervo@​ composite endpoint comprising all the following criteria: risk of 0.2 in a two-­sided test, 183 subjects will be bellvitgehospital.​cat patient alive, stable or with improved quick-­Sequential required in each of the control and experimental groups Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208 1 Open access to obtain statistically significant difference of 12% (considered clinically very uncommon. Nevertheless, because of the risk of the significant). development of resistance when administered as mono- Ethics and dissemination Ethical approval has been obtained from the therapy, fosfomycin must be administered in combination Ethics Committee of Bellvitge University Hospital (AC069/18) and from the with another antibiotic.15 In vitro and small-­scale studies Spanish Medicines and Healthcare Product Regulatory Agency (AEMPS, have demonstrated a synergistic effect of cloxacillin AC069/18), and is valid for all participating centres under existing Spanish plus fosfomycin against S. aureus,16 and several different legislation. The results will be presented at international meetings and will be made available to patients and funders. beta-­lactam combinations have been successfully used in Trial registration number The protocol has been approved by difficult-­to-­treat methicillin-­resistant S. aureus (MRSA) AEMPS with the Trial Registration Number EudraCT 2018-001207-37. ​ infections.17 18 ClinicalTrials.​gov Identifier: NCT03959345; Pre-­results. In a recent multicentre trial, we showed that DAP plus fosfomycin in MRSA bacteraemia achieved better outcomes in a subgroup of younger severely ill patients INTRODUCTION and faster clearance of bacteraemia than DAP alone.19 To Staphylococcus aureus is one of the most common causes of date, however, no other randomised studies evaluating bacteraemia and endocarditis in industrialised countries, the efficacy of cloxacillin plus fosfomycin for treating and has particularly high hospitalisation and mortality MSSA bacteraemia have been published or registered in rates (and associated costs).1 2 Healthcare exposure and the ​ClinicalTrials.​gov database. the increasing use of invasive devices have contributed to We hypothesise that combining cloxacillin plus fosfo- the high burden of the disease.3 mycin during the initial 7 days of treatment achieves Mortality rates at 90 days due to methicillin-­susceptible better outcomes than cloxacillin alone in patients with Staphylococcus aureus (MSSA) bacteraemia range between MSSA bacteraemia. The primary objective of the study is 20% and 30%.4 5 Mortality has been linked to factors such to determine and compare mortality, clinical and micro- as age, comorbidities, source of infection, pathogen viru- biological failure at 7 and 90 days after randomisation by lence elements and optimisation of antibiotic treatment.6 allocated treatment. Complicated S. aureus bacteraemia is common, and is an indicator of poor prognosis.7 Indeed, every continued day of bacteraemia has been associated with a higher risk METHODS AND ANALYSIS of mortality.8 9 Study design and setting Although MSSA bacteraemia is a common and life-­ We will perform a multicentre, superiority, randomised, threatening infection, it is still unclear whether combi- open-­label, phase IV–III, two-­armed parallel group (1:1) nation therapy can reduce duration of bacteraemia or clinical trial. Patients will be recruited from 20 tertiary reduce mortality compared with the current standard of hospitals in Spain (a list of study sites is available in the care (monotherapy beta-­lactams). For over 50 years, the online supplemental material). The trial has been regis- standard treatment of MSSA bacteraemia has been anti- tered in the EudraCT and ClinicalTrials databases. The staphylococcal penicillin monotherapy.10 Today, there is a protocol follows the Standard Protocol Items: Recom- growing interest in identifying new therapeutic regimens mendations for Interventional Trials initiatives, and the able to reduce the rate of therapeutic failure and improve results will be presented in accordance with the Consoli- the outcomes obtained with the standard of care. dated Standards of Reporting Trials statement.20 21 Strategies combining cloxacillin with aminoglyco- sides have not shown any significant improvement in Study population patients’ outcomes, and have been associated with a Inclusion criteria higher risk of nephrotoxicity.11 A randomised multi- ►► Subjects aged ≥18 years. centre study conducted in the UK, which included ►► At least one blood culture positive for MSSA ≤72 hours around 1000 patients and compared the efficacy of the before inclusion, with evidence of active infection. rifampicin combination with the standard treatment for ►► Written informed consent from the participant or the S. aureus bacteraemia, did not show a reduction in early legal representative (LR). or late mortality for the combined therapy compared with monotherapy.12 Nor did two recent studies comparing a Exclusion criteria beta-­lactam and daptomycin (DAP) combination with ►► Severe clinical status with expected death <24 hour. beta-­lactams in monotherapy to treat MSSA bacteraemia ►► Severe hepatic cirrhosis (Child-­Pugh C). show any differences in mortality between groups.4 13 ►► Moderate-­ to-­ severe cardiac chronic failure (NYHA Among the combinations that might improve the (New York Heath Association class III–IV). outcome of patients with MSSA bacteraemia, cloxacillin ►► Prosthetic endocarditis. plus fosfomycin is an appealing strategy. Fosfomycin ►► History of significant allergy to β-lactams or fosfo- is a bactericidal antibiotic which inhibits synthesis of mycin (defined as previous type 1 hypersensitivity N-­acetylmuramic acid, a precursor of bacterial wall pepti- reaction to any β-lactams or fosfomycin, or history doglycan, and is highly active against most strains of S. of serious non-­type 1 hypersensitivity reaction to any aureus.14 Cross-­resistance with other antibiotic groups is penicillin or fosfomycin). 2 Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208 Open access clearance is <30 mL/min, cloxacillin will be adminis- Table 1 Fosfomycin dosage adjusted to renal function tered at dose of 2 g every 6 hours. The fosfomycin dose Creatinine clearance (mL/min) Fosfomycin dosage will be adjusted according to creatinine clearance, as >40 3 g every 6 hours explained in table 1. 20–40 3 g every 12 hours This treatment will be administered during the first 10–20 3 g every 24 hours 7 days after randomisation. After the first week, the choice <10 3 g every 48 hours of antibiotic strategy and the duration of overall antibi- Haemodialysis 3 g after haemodialysis otic treatment will be determined according to clinical criteria by the attending physician, based on current Continuous renal replacement 3 g every 24 hours therapy guidelines. Uncomplicated bacteraemia (no evidence of complicated bacteraemia) will be treated for 10–14 days, and complicated bacteraemia (defined as infection with ►► Known S. aureus fosfomycin non-­susceptibility. haematogenous seeding, progression of infection beyond ►► Polymicrobial bacteraemia with more than one micro-­ the primary focus, persistent bacteraemia, skin alterations organism in blood cultures. suggestive of acute systemic infection, presence of non-­ ►► A positive pregnancy test or pregnancy or lactation at catheter device, haemodialysis) for 4–6 weeks at least, the time of inclusion. depending on the source of the infection and other clin- ►► Myasthenia gravis. ical considerations.22 23 Removal of a focus of infection as ►► Participation in another clinical trial. soon as possible and performance of echocardiogram will ►► Previous participation in the present clinical trial. be prioritised. The assessment schedule is summarised in ►► Social problems, cognitive or psychiatric impairment table 2. A schematic diagram of study design is shown in which might be expected to affect adherence to the figure 1. protocol. ►► Acute SARS-­CoV-2 infection. Outcomes Intervention Efficacy will be analysed by intention to treat in all Patients will be randomly assigned to receive intrave- randomised patients, using a hierarchical testing proce- nous cloxacillin 2 g every 4 hours plus fosfomycin 3 g dure in the following order: treatment success at day 7 every 6 hours, or to receive cloxacillin 2 g every 4 hours followed by treatment success at TOC visit. Furthermore, intravenously for the duration of 7 days. If creatinine a per-­protocol analysis will also be performed. Table 2 The SAFO evaluation schedule Visit day Screening 0 3 7 EOT Unscheduled visit* TOC All patients  Eligibility criteria X  Pregnancy test† X  Informed consent X  Randomisation X  Clinical evaluation X X X X X X‡  Quick SOFA score X X  Blood cultures X X X X X X‡  Blood count and X X X X X‡ biochemical analysis§  Adverse events record X X X X X  Concomitant medication X X X X Subgroup of patients with pharmacokinetic/pharmacodinamic (PK/PD) subanalysis  Lithium heparin blood sample (2×5 mL) X *Unscheduled visit will be performed only in case of clinical infectious symptoms and signs. †Pregnancy test will be performed only in woman of childbearing age. ‡In absence of infective symptoms, clinical assessment may be made by phone call; blood culture and blood analysis will not be necessary. §Complete blood count, biochemical analysis (C reactive protein, creatinine, urea, creatinine clearance, AST, ALT, INR, bilirubin, sodium, potassium, calcium, acid–base analysis) and coagulation test (prothrombin test/INR). ALT, alanine aminotransferase; AST, aspartate amino transferase; EOT, end of treatment; INR, International Normalized Ratio; SOFA, Sequential Organ Failure Assessment; TOC, test of cure. Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208 3 Open access Figure 1 Study design. EOT, end of treatment; TOC, test of cure. Primary endpoints ►► Microbiological relapse, defined by at least one posi- Treatment success at day 7 from randomisation is a tive blood culture for MSSA at least 72 hours after a composite outcome defined by all the following criteria preceding negative culture. met after randomisation: ►► Microbiological treatment failure, defined by a posi- ►► Patient alive at day 7. tive sterile site culture for MSSA at least 14 days after ►► Clinical improvement measured by stable or randomisation. improved quick Sequential Organ Failure Assess- ►► Number of patients with persistent and relapsing ment (SOFA) score (compared with baseline) at day bacteraemia. 7. ►► Number of patients with complicated bacteraemia, ►► Patient afebrile at day 7. defined as persistent bacteraemia, endocarditis or ►► Negative MSSA blood cultures at day 7. metastatic emboli, presence of prosthetic devices. Treatment success at TOC visit, defined by presence of ►► Length of intensive care unit stay. all of the following: ►► Duration of intravenous antibiotic treatment. ►► Patient alive at TOC. We will perform exploratory subgroup analyses for ►► No isolation of MSSA in blood culture and/or at patients at high risk (those with metastatic infection, another sterile site from day 8 until the TOC visit (12 unknown focus of bacteraemia, endocarditis and pneu- weeks after randomisation). In case of patients with a monia) for both primary outcomes. On participants with prolonged course of antibiotic treatment (more than persistent bacteraemia subgroup analysis will be focused 10 weeks), the TOC visit will be performed 2 weeks on treatment success at TOC. after the end of treatment (EOT). Treatment failure is defined by the presence of one of Microbiological the following conditions: all-­cause mortality at TOC, with- ►► In vitro cloxacillin plus fosfomycin combination drawal from the study due to adverse events related to synergy (see online supplemental material). the treatment, requirement of an additional MSSA-­active ►► Emergence of fosfomycin-­ resistant strains during antibiotic until day 7, and lack of clinical improvement therapy in the combination treatment arm. at day 7. ►► Operon agr functionality and its relationship with minimum inhibitory concentration (MIC) changes Secondary endpoints to vancomycin (VAN) and DAP and with biofilm Clinical production. ►► All-­cause mortality at day 7, EOT and TOC visit. ►► VAN and DAP MIC as markers of complications ►► Persistent bacteraemia (at least one positive blood during bacteraemia. Isolates with rising VAN MICs culture) at day 3 and persistent bacteraemia at day 7 are associated with thicker cell walls and dysfunctional after randomisation. agr profiles. These profiles are involved in quorum 4 Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208 Open access sensing, activation of S. aureus toxins and other viru- Allocation lence factors, leading to more resistant but less viru- Participants will be block randomised to receive mono- lent strains. therapy or combination using an internet-­ based, ►► Whole genome sequencing and its changes in patients concealed computer-­ generated random allocation with treatment failure. sequence. Random blocks will be of size 4 or 6. The randomised sequence allocation will be stored in the Pharmacological Biostatistics Unit at Biomedical Research Institute of Bell- Patients recruited at the coordinating centre (Bellvitge vitge (IDIBELL) and will not be available to any member University Hospital) will be included in a pharmacolog- of the research team. ical substudy, after obtaining additional signed informed consent. The variables assessed will be: Data analysis ►► Minimum and maximum concentration in steady The main analysis will be performed for the intention-­ state of fosfomycin and cloxacillin, and pharmacoki- to-­treat population, which will include all randomised netic variability of these concentrations. patients included in the study with a primary outcome ►► Associations between pharmacokinetic parameters assessment. If no statistical significance is detected by day and efficacy. 7 in the hierarchy, then no further hypothesis testing will be performed. The analysis will be repeated in the per-­ Safety protocol population. All patients who receive at least one Safety of cloxacillin plus fosfomycin as compared with dose of treatment will be included in the safety analysis. cloxacillin alone (see online supplemental material). The χ2 test will be used to test the binary endpoints of the success rate. The relative risk for success rate will Follow-up and data collection be calculated, accompanied by 95% CIs. Absolute risk During the first week of treatment, all patients will be difference and 95% CI will also be reported. The time-­ assessed at days 1, 3 and 7 by a member of the investigating to-­event outcomes, including the time of response, and team, and followed up daily by an infectious diseases overall survival will be estimated using the Kaplan-­Meier specialist. Scheduled visits are reported in table 2. A method.24 To account for competing risks, cause-­specific follow-­up visit will be arranged for all participants at EOT cox regression models will be used, and event cause (48 hours after the last dose of antibiotic treatment) and cumulative incidence functions will be plotted.25 All anal- at TOC. At this last visit, a structured telephone interview yses and data management will be performed with R soft- will be performed to assess outcomes. ware, V.4.0.4 or superior.26 All data will be recorded on a secure web applica- tion used for building and managing online databases Monitoring (REDCap). Authorised staff will be free to examine the Monitoring plans records for quality assurance and audit purposes. The data monitoring board will ensure the correct prog- ress of the study in terms of safety, and also the sample Endpoint assessment size assumptions. The primary endpoints will be assessed by a committee comprising three independent senior infectious disease Harms—Data Safety and Monitoring Board (DSMB) specialists with extensive experience in S. aureus bacter- An independent DSMB will review safety data and aemia and endocarditis. This committee will be blinded provide advice about the continuation, modification to treatment allocation and to patient identification. and/or termination of the study, as well as adherence Committee members will receive a data extract containing to the protocol, recruitment, outcomes and additional patients’ demographical data, comorbidities, source of data related to participants’ safety. The DSMB will be infection, quick SOFA score at baseline and day 7, date composed by specialists in pharmacology, biostatistics and results of blood and sterile cultures between rando- and infectious diseases. The review by the DSMB will be misation and TOC, as well as date of death if applicable. performed when half of the sample size will be reached. Statistical analysis plan Adverse events reporting and quantification Sample size An adverse event will be defined as any injury related Prior data indicate a success rate in the cloxacillin alone to medical management occurring during the patient’s group of 74%.4 To achieve a success rate in the experi- participation in the study, even if it is not related to the mental group of 86% (ie, an absolute difference of 12%, study medication. considered as clinically significant), we will need 183 An adverse drug event will be defined as any medication-­ experimental subjects and 183 control subjects to reject related adverse event occurring during the patient’s the null hypothesis of an equal success rate with a prob- participation in the clinical trial. ability of 80%. The probability of type I error associated An adverse drug reaction will be defined as any ‘adverse with this test is 5%, and a dropout rate of 5% has been drug event’ occurring when the medication is used as anticipated. directed and at the usual dosage. Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208 5 Open access Serious adverse event or reaction will be defined as an doubts. Written informed consent will be obtained from event or reaction that: all patients or from their LRs if they lack capacity, before ►► Results in death. enrollment (online supplemental file). Patients (or their ►► Is life-­threatening. LRs) are free to withdraw from the trial at any time; this ►► Causes persistent or significant disability. will be explicitly stated on the patient’s information sheet. ►► Causes a congenital anomaly/birth defect. Patients’ personal and clinical information will be ►► Requires inpatient hospitalisation or prolongation of managed in accordance with European Regulation existing hospitalisation (not related to basal diseases). 2016/679 and Spanish legislation. The trial protocol was approved by the research ethics committee on 28 March Adverse drug events of particular interest for the study 2019 and by the AEMPS on 8 April 2019. The informed ►► Hypokalaemia and hypocalcaemia: blood analysis will consent form and information sheet were approved by be performed every 2–3 days during the first week the research ethics committee on 28 March 2019. The to permit potassium and calcium control. Further- emendation regarding ‘acude SARS-­CoV-2 infection’ as more, administration of potassium supplement will exclusion criteria was approved by the research ethics be recommended from the first day of treatment to committee and by the AEMPS on 29 November 2020. avoid this complication. ►► Sodium overload: since both fosfomycin and cloxa- Data sharing plan and dissemination cillin carry a high sodium load, daily physical exami- Sharing of data generated by this project is an essential nation and administration of a low dose of a diuretic part of our proposed activities and will be carried out in such as furosemide will be recommended to avoid several different ways. We would wish to make our results hypertension, oedema and acute cardiac failure. available both to the community of scientists interested Reporting in infectious diseases and the biology of S. aureus to avoid Any adverse events occurring during the patient’s partici- unintentional duplication of research. pation in the clinical trial will be recorded on the clinical The preliminary results will be presented at interna- chart by the principal investigator (PI) at each scheduled tional and national infectious diseases conferences and visit. The PI will record its possible relationship to the will be published in peer-­reviewed journals. The results study drug. will also be made available to patients, caregivers and The electronic case report form should record only funders through press and social media communications. the following: serious adverse drug events; adverse events A corporative Twitter account will be created to establish (of any degree) related to the study medication, in the direct contact with the general public and other health- opinion of the PI; adverse events (of any degree) leading care professionals. Any formal presentation or publica- to modification of the dosage of the study drug or its tion of data collected from this study will be considered interruption/early discontinuation; adverse events of as a joint publication by the participating investigators particular interest for the study. and will follow the recommendations of the International The sponsor will be notified of all serious adverse events Committee of Medical Journal Editors . within 24 hours of their occurrence. Individual participant data that underlie the results, after deidentification (text, tables, figures and appen- Trial status dices) will be available immediately following publication The SAFO trial opened its first recruitment site on 31 and ending 5 years following article publication. Data will May 2019. The first patient was enrolled on 1 July 2019. be shared with researchers who provide a methodolog- Follow-­up is expected to be completed by May 2022. ically sound proposal to achieve aims in the approved suggestions. Propositions should be directed to the corre- sponding author. DECLARATION Ethics Patients and public involvement The trial will be conducted in accordance with the prin- Patients will not be involved in either the enrollment ciples of the most recent Declaration of Helsinki (agreed or the execution of the trial, or in the assessment of the by the 64th World Medical Association General Assembly interventions. However, before the beginning of the in 2013), the Good Clinical Practice guidelines and the study, a number of patients with previous S. aureus bacte- current local legislation. raemia were contacted by phone to obtain their feedback The study was authorised by the Spanish Medicines about the study. and Healthcare Products Regulatory Agency (AEMPS, 18-0905) and by the Bellvitge University Hospital Ethics Protocol amendments committee (AC069/18). No protocol modifications will become effective until The PI or collaborator at each site will provide patients approved by the relevant authorities and by the Drug with the information sheet, and he/she will explain the Research Ethics Committee (CEIm). Exceptions will be nature of the study and the objectives and clarify any made for any changes to protect patients from imminent 6 Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208 Open access harm and those concerning exclusively logistic or admin- Twitter Oriol Gasch @GaschOriol_Inf istrative aspects. Acknowledgements We thank the CERCA Programme/Generalitat de Catalunya for their institutional support. We would also like to thank the Biostatistics Unit and Author affiliations the Pharmacology department of the IDIBELL for technical help. 1 Department of Infectious Diseases, Hospital Universitari de Bellvitge, L’Hospitalet Collaborators The SAFO study group and the Spanish Network for Research in de Llobregat, Spain Infectious Diseases (REIPI) 2 Bellvitge Institute for Biomedical Research, IDIBELL, Barcelona, Spain Contributors SG, GC, JMA, JC, DB, MAD, AP, SC and MP conceived and designed 3 University of Barcelona, Barcelona, Spain the study. SG, GC, MP and JC wrote and revised the manuscript. CT and NP 4 Department of Infectious Diseases, Hospital Universitario 12 de Octubre, Madrid, designed and wrote statistical analysis plan. PH and SV critically reviewed Spain the protocol. SG-­Z, GC, RS-­J, JMA, LM, JL-­C, OG, AG-­G, SI, GG-­P, EC, LB-­P, IO, 5 Instituto de Investigación Hospital 12 de Octubre, Madrid, Spain AJ-­S, LEL-­C, GE, MA, MJG-­P, FG, JRP, MLP-­B, RMB, MTP-­R, YM, MBL-­Y and GH 6 Complutense University of Madrid, Madrid, Spain contributed to the aquisition of data. All authors have read and approved the final 7 Department of Infectious Diseases, Hospital Clinic de Barcelona, Barcelona, Spain manuscript. 8 Institut d’Investigacions Biomèdiques August Pi i Sunyer, IDIBAPS, Barcelona, Spain Funding The SAFO trial is supported by a competitive grant awarded by the Fondo 9 Department of Infectious Diseases, Consorci Parc de Salut MAR de Barcelona, de Investigaciones Sanitarias at the Spanish government’s National Institute of Barcelona, Spain Health Research, Instituto de Salud Carlos III (ISCIII), (FIS PI17/01116). This study 10 Institut de Recerca Hospital del Mar, IMIM, Barcelona, Spain was supported by Plan Nacional de I+D+i 2017–2021 and Instituto de Salud 11 Department of Infectious diseases, Hospital de la Santa Creu i Sant Pau, Carlos III, Subdirección General de Redes y Centros de Investigación Cooperativa, Barcelona, Spain Ministerio de Economía, Industria y Competitividad, Spanish Network for Research 12 Institut d’Investigació Biomèdica Sant Pau IIB Sant Pau, Barcelona, Spain in Infectious Diseases (REIPI RD16/0016/0005). 13 Infectious Diseases Department, Consorcio Corporacion Sanitaria Parc Tauli, Competing interests GH received a research grant from ERN (19PNJ145). Sabadell, Spain 14 Patient consent for publication Not required. Institut d'Investigació i Innovació Parc Taulí, I3PT, Sabadell, Spain 15 Consorcio Corporación Sanitaria Parc Taulí, Sabadell, Spain Provenance and peer review Not commissioned; externally peer reviewed. 16 Department of Infection and Immunity, Hospital Universitari Sant Joan de Reus, Supplemental material This content has been supplied by the author(s). It has Reus, Spain not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been 17 Departament of Preventive Medicine, Hospital Universitari de Tarragona Joan XXIII, peer-­reviewed. Any opinions or recommendations discussed are solely those Tarragona, Spain of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and 18 Infectious Diseases Unit, Hospital Universitari MutuaTerrassa, Terrassa, Spain responsibility arising from any reliance placed on the content. Where the content 19 Fundació per la Docència i Recerca MútuaTerrassa, Terrassa, Spain includes any translated material, BMJ does not warrant the accuracy and reliability 20 Department of Internal Medicine, Hospital de Sant Joan Despi Moises Broggi, Sant of the translations (including but not limited to local regulations, clinical guidelines, Joan Despi, Spain terminology, drug names and drug dosages), and is not responsible for any error 21 and/or omissions arising from translation and adaptation or otherwise. Territorial Unit of Nosocomial Infection, Hospital Universitari Arnau de Vilanova, Lleida, Spain Open access This is an open access article distributed in accordance with the 22 Institut de Recerca Biomèdica de Lleida, IRBLLEIDA, Lleida, Spain Creative Commons Attribution Non Commercial (CC BY-­NC 4.0) license, which 23 Department of Infectious diseases, Hospital Universitario Virgen Macarena, Seville, permits others to distribute, remix, adapt, build upon this work non-­commercially, Spain and license their derivative works on different terms, provided the original work is 24 Instituto de Biomedicina de Sevilla, Sevilla, Spain properly cited, appropriate credit is given, any changes made indicated, and the use 25 is non-­commercial. See: http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/. Department of Infectious Diseases, Hospital Universitario Cruces, Barakaldo, Spain 26 Biocruces Bizkaia Health Research Institute, Barakaldo, Spain ORCID iDs 27 Microbiology Department, Hospital Universitario Cruces, Barakaldo, Spain Guillermo Cuervo http://​orcid.​org/​0000-​0002-​7075-​943X 28 Internal Medicine, Hospital Universitario Lucus Augusti, Lugo, Spain Oriol Gasch http://​orcid.​org/​0000-​0001-​8518-​458X 29 Instituto de Investigación Sanitaria de Santiago de Compostela, Santiago de Compostela, Spain 30 Department of Infectious diseases, Hospital Universitario Ramon y Cajal, Madrid, Spain 31 Instituto Ramón y Cajal de Investigación Sanitaria, Madrid, Spain REFERENCES 32 Hospital Clínico Universitario Lozano Blesa, Zaragoza, Spain 1 Bergin SP, Holland TL, Fowler VG. Bacteremia, sepsis, and infective 33 endocarditis associated with Staphylococcus aureus. Curr Top Instituto de Investigación Sanitaria Aragón, Zaragoza, Spain 34 Hospital Universitari Germans Trias i Pujol, Badalona, Spain Microbiol Immunol 2015;409:263–96. 35 2 Stewardson AJ, Allignol A, Beyersmann J, et al. The health and Fundació Institut d'Investigació en Ciències de la Salut Germans Trias i Pujol, economic burden of bloodstream infections caused by antimicrobial-­ Badalona, Spain susceptible and non-­susceptible Enterobacteriaceae and 36 Department of Internal Medicine and Infectious Diseases, Hospital Álvaro Staphylococcus aureus in European hospitals, 2010 and 2011: a Cunqueiro, Vigo, Spain multicentre retrospective cohort study. Euro Surveill 2016;21:30319. 37 Hospital de Barcelona, Barcelona, Spain 3 Souli M, Ruffin F, Choi S-­H, et al. Changing characteristics of 38 Staphylococcus aureus bacteremia: results from a 21-­year, Societat Cooperativa d’Instal·lacions Assistencials Sanitàries, Barcelona, Spain prospective, longitudinal study. Clin Infect Dis 2019;69:1868–77. 39 Department of Infectious Diseases, Hospital Universitario La Paz, Madrid, Spain 4 Grillo S, Cuervo G, Carratalà J, et al. Impact of β-lactam and 40 Department of Microbiology and Parassitology, Hospital Universitari de Bellvitge, daptomycin combination therapy on clinical outcomes in methicillin-­ L’Hospitalet de Llobregat, Spain susceptible Staphylococcus aureus bacteremia: a propensity score-­ 41 matched analysis. Clin Infect Dis 2019;69:1480–8. Pharmacy Department, Hospital Universitari de Bellvitge, L’Hospitalet de Llobregat, 5 Rieg S, Joost I, Weiß V, et al. Combination antimicrobial therapy Spain in patients with Staphylococcus aureus bacteraemia-­a post hoc 42 Department of Clinical Pharmacology, Hospital Universitari de Bellvitge, analysis in 964 prospectively evaluated patients. Clin Microbiol Infect L’Hospitalet de Llobregat, Spain 2017;23:406.e1–406.e8. 43 Biostatistics Unit, Institut d’Investigacio Biomedica de Bellvitge, L’Hospitalet de 6 van Hal SJ, Jensen SO, Vaska VL, et al. Predictors of mortality in Staphylococcus aureus bacteremia. Clin Microbiol Rev Llobregat, Spain 44 2012;25:362–86. Biostatistics Unit, Institut d’Investigacio Biomedica de Bellvitge, L'Hospitalet de 7 Gasch O, Camoez M, Dominguez MA, et al. Predictive factors for Llobregat, Spain mortality in patients with methicillin-­resistant Staphylococcus aureus Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208 7 Open access bloodstream infection: impact on outcome of host, microorganism 17 Grabein B, Graninger W, Rodríguez Baño J, et al. Intravenous and therapy. Clin Microbiol Infect 2013;19:1049–57. fosfomycin-­back to the future. Systematic review and meta-­analysis 8 Minejima E, Mai N, Bui N, et al. Defining the breakpoint duration of of the clinical literature. Clin Microbiol Infect 2017;23:363–72. Staphylococcus aureus bacteremia predictive of poor outcomes. Clin 18 del Río A, Gasch O, Moreno A, et al. Efficacy and safety of Infect Dis 2020;70:566–73. fosfomycin plus imipenem as rescue therapy for complicated 9 Kuehl R, Morata L, Boeing C, et al. Defining persistent bacteremia and endocarditis due to methicillin-­resistant Staphylococcus aureus bacteraemia: secondary analysis of a Staphylococcus aureus: a multicenter clinical trial. Clin Infect Dis prospective cohort study. Lancet Infect Dis 2020;20:1409–17. 2014;59:1105–12. 10 Gudiol F, Aguado JM, Almirante B, et al. Executive summary of the 19 Pujol M, Miró J-­M, Shaw E, et al. Daptomycin plus fosfomycin diagnosis and treatment of bacteremia and endocarditis due to versus daptomycin alone for methicillin-­resistant staphylococcus Staphylococcus aureus. A clinical guideline from the Spanish Society bacteremia and endocarditis: a randomized clinical trial. Clin Infect of clinical microbiology and infectious diseases (SEIMC). Enferm Dis 2021;72:1517–25. Infecc Microbiol Clin 2015;33:626–32. 20 Chan A-­W, Tetzlaff JM, Altman DG, et al. SPIRIT 2013 statement: 11 Cosgrove SE, Vigliani GA, Fowler VG, et al. Initial low-­dose defining standard protocol items for clinical trials. Ann Intern Med gentamicin for Staphylococcus aureus bacteremia and endocarditis 2013;158:200–7. is nephrotoxic. Clin Infect Dis 2009;48:713–21. 21 Schulz KFet al. Consort 2010 statement: updated guidelines 12 Thwaites GE, Scarborough M, Szubert A, et al. Adjunctive rifampicin for reporting parallel group randomized trials. Ann Intern Med for Staphylococcus aureus bacteraemia (arrest): a multicentre, 2010;152:726. randomised, double-­blind, placebo-­controlled trial. Lancet 22 Fowler VG, Olsen MK, Corey GR, et al. Clinical identifiers of 2018;391:668–78. complicated Staphylococcus aureus bacteremia. Arch Intern Med 13 Cheng MP, Lawandi A, Butler-­Laporte G, et al. Adjunctive 2003;163:2066–72. daptomycin in the treatment of methicillin-­susceptible 23 Fowler VG, Justice A, Moore C, et al. Risk factors for hematogenous Staphylococcus aureus bacteremia: a randomized, controlled trial. complications of intravascular catheter-­associated Staphylococcus Clin Infect Dis 2021;72:e196–203. aureus bacteremia. Clin Infect Dis 2005;40:695–703. 14 Popovic M, Steinort D, Pillai S, et al. Fosfomycin: an old, new friend 24 Kaplan EL, Meier P. Nonparametric estimation from incomplete Eur J Clin Microbiol Infect Dis 2010;29:127–42. observations. J Am Stat Assoc 1958;53:457–81. 15 Drugeon HB, Courtieu AL. The role of culture media on the 25 Wolkewitz M, Cooper BS, Bonten MJM, et al. Interpreting and fosfomycin sensitivity of six Serratia strains and their resistant comparing risks in the presence of competing events. BMJ mutants. Chemotherapy 1982;28:345–50. 2014;349:g5060. 16 Kastoris AC, Rafailidis PI, Vouloumanou EK, et al. Synergy of 26 R core team. R: a language and environment for statistical fosfomycin with other antibiotics for gram-­positive and gram-­ computing. Austria: R Found Stat Comput Vienna, 2016. http//​ negative bacteria. Eur J Clin Pharmacol 2010;66:359–68. wwwR-​Project.​org 8 Grillo S, et al. BMJ Open 2021;11:e051208. doi:10.1136/bmjopen-2021-051208