Articles

Efficacy and safety of an early oral switch in low-risk

Staphylococcus aureus bloodstream infection (SABATO):
an international, open-label, parallel-group, randomised,
controlled, non-inferiority trial
Achim J Kaasch, Luis Eduardo López-Cortés, Jesús Rodríguez-Baño, José Miguel Cisneros, M Dolores Navarro, Gerd Fätkenheuer, Norma Jung,
Siegbert Rieg, Raphaël Lepeule, Laetitia Coutte, Louis Bernard, Adrien Lemaignen, Katrin Kösters, Colin R MacKenzie, Alex Soriano, Stefan Hagel,
Bruno Fantin, Matthieu Lafaurie, Jean-Philippe Talarmin, Aurélien Dinh, Thomas Guimard, David Boutoille, Tobias Welte, Stefan Reuter,
Jan Kluytmans, Maria Luisa Martin, Emmanuel Forestier, Hartmut Stocker, Virginie Vitrat, Pierre Tattevin, Anna Rommerskirchen, Marion Noret,
Anne Adams, Winfried V Kern, Martin Hellmich, Harald Seifert, for the SABATO study group*

Summary
Background Staphylococcus aureus bloodstream infection is treated with at least 14 days of intravenous antimicrobials. Lancet Infect Dis 2024
We assessed the efficacy and safety of an early switch to oral therapy in patients at low risk for complications related Published Online
to S aureus bloodstream infection. January 17, 2024
https://doi.org/10.1016/
S1473-3099(23)00756-9
Methods In this international, open-label, randomised, controlled, non-inferiority trial done in 31 tertiary care
See Online/Comment
hospitals in Germany, France, the Netherlands, and Spain, adult patients with low-risk S aureus bloodstream infection https://doi.org/10.1016/
were randomly assigned after 5–7 days of intravenous antimicrobial therapy to oral antimicrobial therapy or to S1473-3099(24)00032-X
continue intravenous standard therapy. Randomisation was done via a central web-based system, using permuted *Members listed at the end of
blocks of varying length, and stratified by study centre. The main exclusion criteria were signs and symptoms of the Article
complicated S aureus bloodstream infection, non-removable foreign devices, and severe comorbidity. The composite For the German translation of
primary endpoint was the occurrence of any complication related to S aureus bloodstream infection (relapsing the abstract see Online for
appendix 1
S aureus bloodstream infection, deep-seated infection, and mortality attributable to infection) within 90 days, assessed
For the Spanish translation of the
in the intention-to-treat population by clinical assessors who were masked to treatment assignment. Adverse events
abstract see Online for
were assessed in all participants who received at least one dose of study medication (safety population). Due to slow appendix 2
recruitment, the scientific advisory committee decided on Jan 15, 2018, to stop the trial after 215 participants were For the French translation of the
randomly assigned (planned sample size was 430 participants) and to convert the planned interim analysis into the abstract see Online for
final analysis. The decision was taken without knowledge of outcome data, at a time when 126 participants were appendix 3
enrolled. The new sample size accommodated a non-inferiority margin of 10%; to claim non-inferiority, the upper For the Dutch translation of the
bound of the 95% CI for the treatment difference (stratified by centre) had to be below 10 percentage points. The trial abstract see Online for
appendix 4
is closed to recruitment and is registered with ClinicalTrials.gov (NCT01792804), the German Clinical trials register
Institute of Medical
(DRKS00004741), and EudraCT (2013–000577–77). Microbiology and Hospital
Hygiene, Medical Faculty,
Findings Of 5063 patients with S aureus bloodstream infection assessed for eligibility, 213 were randomly assigned to Otto von Guericke University
switch to oral therapy (n=108) or to continue intravenous therapy (n=105). Mean age was 63·5 (SD 17·2) years and Magdeburg, Magdeburg,
Germany (Prof A J Kaasch MD);
148 (69%) participants were male and 65 (31%) were female. In the oral switch group, 14 (13%) participants met the Unidad Clínica de
primary endpoint versus 13 (12%) in the intravenous group, with a treatment difference of 0·7 percentage points Enfermedades Infecciosas y
(95% CI –7·8 to 9·1; p=0·013). In the oral switch group, 36 (34%) of 107 participants in the safety population had at Microbiología, Hospital
least one serious adverse event compared with 27 (26%) of 103 participants in the intravenous group (p=0·29). Universitario Virgen Macarena
(L E López-Cortés MD,
Prof J Rodríguez-Baño MD) and
Interpretation Oral switch antimicrobial therapy was non-inferior to intravenous standard therapy in participants Unidad Clínica de
with low-risk S aureus bloodstream infection. However, it is necessary to carefully assess patients for signs and Enfermedades Infecciosas,
Microbiología y Parasitología,
symptoms of complicated S aureus bloodstream infection at the time of presentation and thereafter before considering
Hospital Universitario Virgen
early oral switch therapy. del Rocío (Prof J M Cisneros MD,
M Dolores Navarro MD),
Funding Deutsche Forschungsgemeinschaft. Instituto de Biomedicina de
Sevilla (IBiS)/CSIC, Department
of Medicine, Universidad de
Copyright © 2024 Elsevier Ltd. All rights reserved. Sevilla, Seville, Spain; Centro de
Investigación Biomédica en
Introduction metastatic complications, and relapse are common.3 Red de Enfermedades
Infecciosas (CIBERINFEC),
Staphylococcus aureus bloodstream infection affects Low-risk S aureus bloodstream infection occurs in a
Instituto de Salud Carlos III,
20–30 people per 100 000 population annually and has a subset of patients in whom risk factors and the clinical Madrid, Spain (L E López-Cortés,
3-month mortality of 20–30%.1,2 Metastatic foci, late course do not suggest elevated risk for the presence or Prof J Rodríguez-Baño,

www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9 1

Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
 Articles
Prof J M Cisneros,
M Dolores Navarro);
Department I of Internal
Medicine, University Clinics
(Prof G Fätkenheuer MD,
N Jung MD), Institute of
Medical Statistics and
Computational Biology
(A Adams MSc,
Prof M Hellmich PhD), and
Institute for Medical
Microbiology, Immunology
and Hygiene
(Prof H Seifert MD), Faculty of
Medicine, University Hospital
Cologne, University of Cologne,
Cologne, Germany; Division of
Infectious Diseases,
Department of Medicine II,
Faculty of Medicine, University
Medical Centre Freiburg,
Freiburg, Germany
(Prof S Rieg MD,
Prof W V Kern MD);
Antimicrobial Stewardship
Team, Department of
Prevention, Diagnosis, and
Treatment of Infections, Henri-
Mondor University Hospital,
Creteil, France (R Lepeule MD,
L Coutte MD); Service de
Médecine Interne et Maladies
Infectieuses, Centre Hospitalier
Régional Universitaire de Tours,
Tours, France
(Prof L Bernard MD,
A Lemaignen MD); Medical
Clinic II—Clinic for
Gastroenterology, Hepatology,
Neurogastroenterology,
Infectious Diseases,
Hematology, Oncology and
Palliative Medicine, Helios
Klinikum Krefeld, Krefeld,
Germany (K Kösters MD);
Institute of Medical
Microbiology and Hospital
Hygiene, University Hospital
Düsseldorf, Düsseldorf,
Germany
(Prof C R MacKenzie MD,
A Rommerskirchen MSc);
Department of Infectious
Diseases, Hospital Clínic,
Institut d’Investigacions
Biomèdiques Agustí Pi i Sunyer
(IDIBAPS), CIBERINFEC,
Barcelona, Spain
(A Soriano MD PhD); Institute
for Infectious Diseases and
Infection Control, Jena
University Hospital—Friedrich
Schiller University Jena, Jena,
Germany (S Hagel MD); Internal
Medicine Department, Hôpital
Beaujon, Assistance Publique
Hôpitaux de Paris, Clichy,
France (Prof B Fantin MD);
Infectious Diseases
Department, Saint-Louis
Hospital, Paris, France
2 www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9
Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
Research in context
Evidence before this study
The recommended treatment for Staphylococcus aureus
bloodstream infection is at least 14 days of antimicrobial
therapy, which is usually administered intravenously. It is
unclear whether an early switch to oral antimicrobials is safe
and effective. We searched PubMed for studies in adult patients
with S aureus bloodstream infection from database inception to
Nov 10, 2023, using the terms (“oral” OR “per os”) AND
“Staphylococcus aureus” AND (“bacteraemia” OR “bloodstream
infection”). Several retrospective observational studies found a
similar or lower mortality in patients with S aureus bloodstream
infection receiving early oral antimicrobial switch therapy than
standard intravenous therapy. However, since these studies are
not randomised, they are prone to bias, particularly
confounding by indication. Metastatic foci, relapse rate, and
late metastatic complications are more common in S aureus
bloodstream infection than in other bloodstream infections.
Therefore, studies addressing severe infections with other
microorganisms cannot be extrapolated to S aureus.
A systematic review identified 21 randomised controlled trials,
published until March, 2021, comparing oral with intravenous
treatment for osteomyelitis, bacteraemia, and endocarditis.
No significant difference was found for overall treatment
success in ten trials assessing bacteraemia (risk difference 0·07,
95% CI 0·01–0·15) and four trials on infective endocarditis
development of complicated infection (ie, deep-seated
infection, prolonged bacteraemia, and metastatic foci).4–6
In a prospective cohort study conducted in 2006–11,
292 (23%) of 1288 patients with S aureus bloodstream
infection met criteria for low-risk infection.4
The optimal antimicrobial therapy in patients with
S aureus bloodstream infection has been a topic of
considerable debate. Guidelines recommend at least
14 days of intravenous antimicrobial therapy for low-risk
S aureus bloodstream infection,7,8 whereas the updated
UK guidelines for the treatment of meticillin-resistant
S aureus (MRSA) bacteraemia consider co-trimoxazole as
an oral step-down therapy.9 An early switch to oral
antimicrobial therapy might reduce hospital stays and
infusion-related complications. However, S aureus
bloodstream infection might not be treated successfully
if adequate serum concentrations are not reached by oral
medication. Furthermore, earlier outpatient management
could lead to reduced compliance with the prescribed
drug regimen and delayed identification of complications.
Randomised controlled trials directly addressing an
early oral switch in S aureus bloodstream infection are
scarce, and previous trials have included few participants
with S aureus bloodstream infection receiving oral
therapy.10 Two trials have shown non-inferiority of an oral
switch in patients with endocarditis11 and bone and joint
infection.12 However, the first trial mainly included
participants with organisms other than S aureus, and in
(0·08, 0·03–0·14). However, none of these studies focused on
S aureus bloodstream infection.
Added value of this study
This study is the first randomised controlled trial to show non-
inferiority of early oral antimicrobial switch therapy compared
with standard intravenous treatment in patients with low-risk
S aureus bloodstream infection. The findings support results
from observational studies that reported similar outcomes
between groups.
Implications of all the available evidence
The trial indicates that in the subgroup of patients with low-risk
S aureus bloodstream infection, a switch to oral antimicrobial
medication can be considered after 5–7 days of appropriate
anti-staphylococcal intravenous therapy. Patients with low-risk
S aureus bloodstream infection represent a small subgroup of
patients with S aureus bloodstream infection, and a careful
assessment of the individual patient is a prerequisite for
assigning the low-risk category. Our study included only a few
patients with meticillin-resistant S aureus (MRSA), which might
limit generalisability of the results to patients with bloodstream
infection caused by MRSA. Future studies should address other
subgroups of patients with S aureus bloodstream infection,
particularly from other risk categories, and confirm the benefits
of early oral switch.
the second trial, patients with S aureus bloodstream
infection were excluded. Retrospective observational
studies have found similar or lower 30-day or 90-day
mortality for oral switch therapy compared with
intravenous therapy in patients with S aureus bloodstream
infection.5,6,13–19 However, retrospective studies have a high
risk of confounding by indication. Two surveys reported
that less than 20% of infectious disease physicians from
the USA and Canada felt comfortable with switching to
oral medication in patients with S aureus bloodstream
infection.20,21
Here, we report the results of the SABATO trial, a
randomised non-inferiority trial, which aimed to evaluate
the efficacy and safety of an early switch to oral
antimicrobial therapy in patients with low-risk S aureus
bloodstream infection compared with standard
intravenous antimicrobial therapy.
Methods
Study design
We conducted an international, open-label, parallel-
group, randomised, controlled, non-inferiority trial in
patients with low-risk S aureus bloodstream infection in
31 tertiary care hospitals in four European countries
(Germany, France, the Netherlands, and Spain; figure 1;
appendix 5 pp 11–12). The non-inferiority design was
justified by expected benefits for patients on early oral
switch therapy, such as a lower rate of complications
 Articles

(M Lafaurie MD); Department of

5063 participants assessed for eligibility Infectious Diseases, Centre
Hospitalier de Cornouaille,
Quimper, France
4850 ineligible (J-P Talarmin MD); Infectious
4694 did not meet eligibility criteria Diseases Department,
154 were not asked to be randomised or declined to participate Raymond-Poincaré University
2 randomised in error Hospital, Garches, France
(A Dinh MD); Infectious
Diseases Department, CHD
213 randomised (intention-to-treat Vendée, La Roche-sur-Yon,
population) France (T Guimard MD);
Department of Infectious
Diseases, University Hospital of
Nantes and CIC 1413, INSERM,
108 assigned to the oral switch group 105 assigned to the intravenous group Nantes, France
(Prof D Boutoille MD PhD); Clinic
for Respiratory Medicine and
1 discontinued (positive follow-up blood culture) 2 discontinued (withdrew informed consent) Infectious Diseases, Member of
the German Center of Lung
Research, Medical School
Hannover, Hannover, Germany
107 received at least one dose of study drug and 103 received at least one dose of study drug and
included in safety population included in safety population (Prof T Welte MD); Department
of Infectious Diseases and
General Internal Medicine,
Department of Infection
2 excluded (insufficient duration of pre-study antimicrobial 2 excluded
therapy) 1 positive follow-up blood culture before randomisation Control, Klinikum Leverkusen,
1 intravascular prosthesis Leverkusen, Germany
(Prof S Reuter MD); Department
of Medical Microbiology,
105 included in modified intention-to-treat 101 included in modified intention-to-treat University Medical Center
population population Utrecht, Utrecht University,
Utrecht, Netherlands
(Prof J Kluytmans MD);
19 excluded 22 excluded Infectious Diseases Unit,
17 protocol violations* 23 protocol violations* Internal Medicine Department,
2 withdrew consent after starting study drug 1 missing data on attributability of death Hospital Universitari Son
2 missing data on attributability of death Espases, Fundació Institut
d’Investigació Sanitària Illes
Balears, Palma de Mallorca,
86 included in clinically evaluable population 79 included in clinically evaluable population Spain (M L Martin MD);
Infectious Diseases
Department, Centre Hospitalier
Figure 1: Trial profile
Métropole Savoie, Chambéry,
Four participants in the oral group and three in the intravenous group were lost to follow-up. *Some participants had more than one protocol violation.
France (E Forestier MD); Klinik
für Infektiologie, St Joseph
Hospital Berlin Tempelhof,
related to intravenous therapy and a shorter hospital stay. present before enrolment, such as a deep-seated focus (eg, Berlin, Germany (H Stocker MD);
Although the study completed in December, 2019, data endocarditis, pneumonia, infected implant, undrained Infectious Diseases Unit, Centre
analysis and sharing were delayed by the COVID-19 abscess, empyema, and osteomyelitis), septic shock within Hospitalier d’Annecy Genevois,
pandemic. The study protocol was published and 4 days before random assignment,24 prolonged Epagny Metz-Tessy, France
(V Vitrat MD); Infectious
approved by the ethics committees of all participating bacteraemia (defined as a positive blood culture obtained Diseases and Intensive Care
centres and the respective competent authorities.22,23 An >72 h after start of adequate antimicrobial therapy, even if Unit, Pontchaillou University
independent data monitoring committee periodically there was a negative follow-up blood culture thereafter), Hospital, Rennes, France
assessed the safety of the trial participants. and a body temperature higher than 38°C measured on (Prof P Tattevin MD); French
National Network of Clinical
both days before random assignment. Patients were Research in Infectious Diseases,
Participants excluded if intravascular catheters were not removed Assistance Publique-Hôpitaux
Adult patients (aged ≥18 years) with S aureus isolated from within 4 days after the first positive blood culture was de Paris, Paris, France
at least one blood culture were eligible if they had received drawn. Patients with a higher a-priori risk for (M Noret MSc); German Centre
for Infection Research, Partner
5–7 days of appropriate intravenous antimicrobial therapy, complications related to S aureus bloodstream infection Site Bonn-Cologne, Cologne,
initiated within 72 h after the first positive blood culture were also excluded—namely, patients with a recent history Germany (Prof H Seifert);
was drawn, and at least one follow-up blood culture of S aureus bloodstream infection within the preceding Institute of Translational
Research, CECAD Cluster of
obtained within 24–96 h after the start of appropriate 3 months, injection drug use, severe immunodeficiency
Excellence, University of
antimicrobial therapy. Blood cultures taken in this period or severe immunosuppression, or presence of a prosthetic Cologne, Cologne, Germany
had to be negative for S aureus for the patient to be heart valve or deep-seated vascular graft. (Prof H Seifert)
included. Patients were excluded if signs and symptoms After a protocol amendment (ethics approval on
of complicated S aureus bloodstream infection were Dec 29, 2014), patients with end-stage renal disease,

www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9 3

Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
 Articles
Correspondence to:
Prof Achim J Kaasch, Institute of
Medical Microbiology and
Hospital Hygiene, Medical
Faculty, Otto von Guericke
University Magdeburg,
39120 Magdeburg, Germany
achim.kaasch@med.ovgu.de
See Online for appendix 5
4 www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9
Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
severe liver disease, a prosthetic joint, or a pacemaker
were eligible if the infective focus was a removable
intravascular catheter or a skin or soft-tissue infection.23
Prosthetic joints and pacemakers needed to be implanted
more than 6 months before enrolment. Patients with
end-stage renal disease or a pacemaker required an
echocardiogram to rule out endocarditis (appendix 5
p 25). All participants gave written informed consent
before enrolment.
Randomisation and masking
Participants were randomly allocated to treatment (1:1) no
earlier than 1 day before starting the study drug through a
central internet randomisation service (TENALEA),
stratified by study centre with permuted blocks of varying
length. Participants and investigators were unmasked to
treatment assignment. An independent clinical review
committee, who made the final adjudication, was masked
to group assignment (drug names were removed and
only masked information was provided to the committee).
Procedures
Participants either switched to oral therapy or continued
intravenous standard therapy after 5–7 days of
intravenous standard therapy, with a total duration of
antimicrobial therapy of 14 days. Oral antimicrobials
were selected by the study physician according to
susceptibility results and suspected allergy or intolerance
in the following order: co-trimoxazole (160 mg
trimethoprim plus 800 mg sulfamethoxazole) every 12 h
for meticillin-susceptible S aureus (MSSA) and MRSA,
then clindamycin 600 mg every 8 h for MSSA or linezolid
600 mg every 12 h for MRSA in the oral switch group;
and intravenous flucloxacillin 2 g every 6 h (cloxacillin
2 g every 6 h in Spain and France), cefazolin 2 g every 8 h,
or vancomycin 1 g every 12 h for MSSA or vancomycin 1 g
every 12 h or daptomycin 6–10 mg/kg every 24 h for
MRSA in the intravenous standard therapy group. Dose
adjustments in individual patients were performed as
judged appropriate by the local investigator. Therapeutic
drug monitoring for vancomycin was recommended
to be performed once weekly starting before the
fourth vancomycin dose, and the target vancomycin
trough concentration was 10–20 µg/mL. The local
hospital pharmacy provided the study drugs and standard
drugs from the pharmacy were used.
Data on participants’ sex were collected by the study
teams and defined as recorded in the hospital information
system. Ethnicity data were not collected.
Follow-up started at the end of therapy and ended
90 days from the first positive blood culture. A 30-day
follow-up visit was added as a protocol amendment
(Dec 29, 2014). Antimicrobial treatment, S aureus
bloodstream-related complications, length of admission
to hospital, survival, adverse events, Clostridium difficile
infection, and complications of intravenous therapy were
assessed during the follow-up period and during
follow-up visits. Adverse events were graded according to
the Common Terminology Criteria for Adverse Events,
version 4.0.25 Only events graded 3 or worse were
recorded and reported from the start of treatment to the
end of the study. C difficile infection was determined via a
positive test result during a diarrhoea episode.
Outcomes
The composite primary endpoint was complications
related to S aureus bloodstream infection within 90 days,
defined as relapsing S aureus bloodstream infection,
deep-seated infection with S aureus, or death attributable
to S aureus bloodstream infection. This outcome was
centrally assessed. Relapsing S aureus bloodstream
infection was defined as a positive blood culture for
S aureus within the intervention or follow-up period.
Deep-seated infection was defined as any deep-seated
focus of S aureus infection resulting from haematogenous
dissemination. Diagnosis of deep-seated infection
required a positive culture from the respective site or a
blood culture positive with S aureus plus imaging studies
indicating a deep-seated focus. Death was attributed to
S aureus infection when at least one of the following
conditions was present: positive blood culture for
S aureus taken within 72 h before death, persistent focus
of deep-seated S aureus infection at time of death,
persistent signs and symptoms of systemic infection at
time of death as judged by the study physician, or post-
mortem analysis proving S aureus-related complication
as the cause of death. All other deaths were classified as
unrelated to S aureus.
Secondary outcomes were the length of hospital stay
from the date of first positive blood culture to hospital
discharge or death; complications of intravenous
administration (caused by study medication or not) from
the start of study drug to the end of follow-up; C difficile
infection; 14-day, 30-day, and up to 90-day survival;
adverse events; and serious adverse events.
Statistical analysis
The trial was originally designed with a 5% non-
inferiority margin, which yielded a total sample size of
430 participants (one-sided α=0·05, β=0·2; one interim
analysis at an information fraction of 0·5 [ie,
215 participants] using the O’Brien-Fleming bound
of 2·373), with 2·5% of participants reaching the
endpoint plus adjustments for deaths unrelated to
S aureus blood infection (10%), protocol violations (10%),
and stratification (5%; appendix 5 pp 5–6). To
accommodate conflicting views in the review board of
the German Research Foundation about the size of the
non-inferiority margin (ie, 5% or 10%) we planned a
sequentially rejective testing procedure (hierarchical
testing stratified by centre) based on Zhao’s test from
weaker to stronger assertions.26 Due to slow recruitment,
the scientific advisory committee decided on Jan 15, 2018,
to convert the interim analysis to the final analysis. The

decision was taken without knowledge of outcome data,
at a time when 126 participants were enrolled. The new
sample size still accommodated a 10% non-inferiority
margin (one-sided α=0·05, β=0·2, expected event rate
of 2·5%, hierarchical testing stratified by centre with
Zhao’s test).
According to recommendations from the European
Medicines Agency, we initially selected the clinically
evaluable population (ie, per-protocol population) for the
primary analysis and the intention-to-treat population for
the coprimary analysis.27,28 However, updated guidance
recommends the intention-to-treat population for the
primary analysis,29 and we therefore report both
populations side by side, with a focus on the intention-to-
treat population.
We also assessed the primary endpoint in the modified
intention-to-treat population, which comprised parti­
cipants who were randomly assigned and received any
study drug, excluding participants in whom an inclusion
criterion incompatible with low-risk S aureus bloodstream
infection was violated before starting the study drug. For
example, participants were excluded from the modified
intention-to-treat population when follow-up blood
cultures (obtained before the start of study drug) turned
positive after the start of study drug. In both intention-to-
treat populations, data were analysed as assigned, with
indeterminate and missing outcomes counted as the
primary outcome being met.
The clinically evaluable population included all
participants who were treated according to the protocol
with study drug, were observed until the end of follow-
up, or they reached the primary endpoint and did not
receive antimicrobial therapy during follow-up that could
have masked the primary endpoint. Study drug was
considered according to the protocol when at least 5 days
of initial intravenous therapy was given and the total
duration of antimicrobial therapy was 12–16 days, thus
allowing a variation of 15% around the 14-day target
duration. The safety population included all participants
who received any study drug. Specifically, participants
who ever received an oral antimicrobial agent were
compared with participants who never received an oral
antimicrobial agent.
Prespecified subgroups with at least ten events were
formed according to baseline data of age, sex, meticillin
susceptibility, focus of infection, specific comorbidities
(pacemaker, prosthetic joint, moderate or severe liver
disease, end-stage renal disease, and immuno­
suppression), Charlson Comorbidity Index, country, and
whether any echocardiography was performed.
In sensitivity analyses, we estimated the treatment
difference using alternative definitions for endpoints in
the intention-to-treat, modified intention-to-treat, and
clinically evaluable populations. We did sensitivity analyses
wherein: (1) missing or indeterminate outcomes were not
considered to indicate that the primary endpoint had been
met (prespecified analysis); (2) only complications related
www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9
Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
Articles
to S aureus bloodstream infection that occurred more than
7 days after starting the study drug were counted as
meeting the primary endpoint and missing or
indeterminate outcomes were disregarded (thus,
complications that could have been identified with a more
stringent work-up when starting the study drug were
excluded; post-hoc analysis); (3) only complications related
to S aureus bloodstream infection that were confirmed
with a microbiological result were counted as events and
missing or indeterminate outcomes were disregarded
(prespecified analysis); and (4) the primary endpoint was
modified to include all-cause mortality (prespecified
analysis).
The difference in median lengths of stay between
MSSA and MRSA was evaluated by the Mann-Whitney
U test and corresponding Hodges-Lehman 95% CI.
Kaplan-Meier curves were calculated to assess differences
in time to S aureus bloodstream infection-related
complications and the length of hospital stay (prespecified
analyses) and compared with an unstratified log-rank
test. For time-to-event analyses, hazard ratios were
estimated from Cox regression. To assess the effect of
non-attributable mortality, competing risks were
calculated according to Fine and Gray, considering non-
attributable mortality as a competing event (post-hoc
sensitivity analysis).30 Safety endpoints were assessed
with Fisher’s exact test. Rates and rate ratios of adverse
events were calculated by Poisson regression. The p value
for the primary endpoint is for non-inferiority; all other
p values are for superiority. Statistical calculations were
done using SAS (version 9.4) and Stata (version 17.0).
Reporting of the trial followed the extension of the For CONSORT extensions see
CONSORT statement to non-inferiority trials. A masked, https://www.equator-network.
org/reporting-guidelines/
independent clinical review committee carefully
consort/
evaluated study outcomes and treatment. This trial is
registered with ClinicalTrials.gov (NCT01792804), the
German clinical trials register (DRKS00004741), and
EudraCT (2013–000577–77).
Role of the funding source
The funder of the study had no role in study design, data
collection, data analysis, data interpretation, or writing of
the report.
Results
From Dec 20, 2013, to Dec 22, 2019, 5063 participants
were assessed for eligibility and 213 were enrolled in the
intention-to-treat population (figure 1; appendix 5
pp 17–19). 108 participants were assigned to the oral
switch group and 105 to the intravenous group.
The clinically evaluable population consisted of
165 participants, with 86 participants in the oral switch
group and 79 participants in the intravenous standard
therapy group. The safety population comprised
210 participants who received at least one dose of study
drug, including 107 participants in the oral switch group
and 103 in the intravenous group.
5
 Articles

In the intention-to-treat population, baseline charac­ infection. There were numerically more participants in
teristics were largely balanced between groups and the oral switch group than the intravenous group with
participants were overall representative of low-risk moderate or severe liver disease (11 [10%] of 108 vs
S aureus infection (table 1). Mean age was 63·5 years four [4%] of 105) and diabetes (44 [41%] vs 28 [27%]).
(SD 17·5), and 148 (69%) participants were male and Before starting study medication, participants received a
65 (31%) were female. Overall, catheter-related infection median of 6 days of intravenous antimicrobials in both
and skin and soft-tissue infection were the most prevalent groups. Baseline characteristics in the clinically evaluable
foci, and 16 (8%) participants had an MRSA bloodstream population were similar to the intention-to-treat
population.
Intention-to-treat population Clinically evaluable Participants in the intention-to-treat population were
population treated with study medication for a median duration of
Oral switch Intravenous Oral switch Intravenous 8 days (IQR 7–9) in both groups, resulting in a full
group (n=108) group (n=105) group (n=86) group (n=79) treatment course with antimicrobials of a median of
Age, years 64·4 (16·8) 62·6 (17·6) 62·4 (17·1) 62·1 (17·8) 14 days (IQR 14–15; appendix 5 p 20). Of 108 participants
Sex in the oral switch group, 63 (58%) received co-
Male 71 (66%) 77 (73%) 59 (69%) 58 (73%) trimoxazole, 35 (32%) received clindamycin, nine (8%)
Female 37 (34%) 28 (27%) 27 (31%) 21 (27%) received linezolid, and one (1%) did not receive any study
BMI, kg/m2 27·6 (6·7) 25·6 (5·4) 27·4 (6·5) 26·1 (5·3)
medication. Of 105 participants in the intravenous group,
Time in hospital before 10 (7–14) 11 (7–15) 10 (7–13) 10 (7–13)
46 (44%) received cefazolin, 45 (43%) received intravenous
randomisation, days flucloxacillin or cloxacillin, seven (7%) received
Intervention to remove or drain 9 (8%) 14 (13%) 5 (6%) 10 (13%) vancomycin, five (5%) received daptomycin, and two (2%)
infective focus did not receive any study medication. Nine (4%) of
Echocardiography (TTE or TEE, or 69 (64%) 60 (57%); 1 55 (64%) 46 (58%) 213 participants who initially started study medication
both) performed within 7 days before switched to an alternative study medication (four oral
or after randomisation
switch and five intravenous standard therapy).
CRP at baseline visit, mg/L* 35 (13–70); 13 23 (10–59); 15 35 (14–64); 11 20 (11–52); 13
In the intention-to-treat population, the primary
Resistance of Staphylococcus aureus isolate
composite outcome (ie, complication related to S aureus
Meticillin 6 (6%) 10 (10%) 4 (5%) 5 (6%)
bloodstream infection or a missing outcome within
Clindamycin 12 (12%); 4 11 (11%); 5 10 (12%); 4 7 (10%); 5
90 days) occurred in 14 (13%) of 108 participants in the
Co-trimoxazole 1 (1%); 2 3 (3%); 1 0 (0%); 2 2 (3%)
oral switch group and 13 (12%) of 105 in the intravenous
Focus of infection
standard therapy group (table 2). The treatment
Peripheral venous catheter 47 (44%) 46 (44%) 41 (48%) 35 (44%) difference was 0·7 percentage points (95% CI –7·8 to 9·1;
Central venous catheter 24 (22%) 25 (24%) 18 (21%) 16 (20%) p=0·013). In the clinically evaluable population,
Skin and soft-tissue infection 26 (24%) 22 (21%) 19 (22%) 19 (24%) complications related to S aureus bloodstream infection
Other† 5 (5%) 4 (4%) 4 (5%) 3 (4%) occurred in three (4%) of 86 participants in the oral
Not identified 6 (6%) 8 (8%) 4 (5%) 6 (8%) switch group versus four (5%) of 79 participants in the
Comorbidities intravenous group (treatment difference –2·9 percentage
Moderate or severe liver disease 11 (10%) 4 (4%) 8 (10%) 2 (3%) points, 95% CI –9·6 to 3·9; p<0·0001). Non-inferiority
Chronic renal failure 17 (16%); 1 18 (17%) 10 (12%); 1 12 (15%) was achieved in both populations (upper bound of the
End-stage renal disease 9 (8%); 1 5 (5%) 5 (6%); 1 4 (5%) 95% CI was <10 percentage points for the intention-to-
Chronic lung disease 14 (13%); 1 17 (16%) 10 (12%) 11 (14%) treat population and <5 percentage points for the
Diabetes without end-organ 25 (23%) 18 (17%) 21 (24%) 11 (14%) clinically evaluable population). The results showed
damage stability in sensitivity analyses when addressing missing
Diabetes with end-organ damage 19 (18%) 10 (10%) 13 (15%) 9 (11%) or indeterminate outcomes, the contribution of
Any immunosuppression‡ 16 (15%) 16 (15%); 1 12 (14%) 14 (18%); 1 complications related to S aureus bloodstream infection
Charlson Comorbidity Index 3 (1–5) 3 (1–4) 2 (1–4) 3 (1–4) that occurred more than 7 days after random assignment,
Intravenous antimicrobials before 6 (6–7) 6 (5–7) 7 (6–7) 6 (5–7) and whether complications were microbiologically
randomisation, days
confirmed as being related to S aureus infection
Data are mean (SD), n (%), median (IQR), n (%); missing, or median (IQR); missing. TEE=transoesophageal (appendix 5 pp 13–14). However, when non-attributable
echocardiography. TTE=transthoracic echocardiography. *Laboratory results were from the day of randomisation;
if unavailable, the closest result within 3 days before the start of study drug was accepted. †In the intention-to-treat
mortality was included in the primary endpoint, the
population, the sites of other focus were urinary tract (n=2), suspected pulmonary focus (n=2), and cholecystitis (n=1) 95% CI crossed the 10% non-inferiority margin in the
in the oral switch group and urinary tract (n=4) in the intravenous group. ‡Any immunosuppression comprises intention-to-treat population but not in the clinically
systemic corticosteroid therapy for more than 1 week, current antineoplastic chemotherapy, any other
evaluable population (appendix 5 pp 13–14). Cumulative
immunosuppressive therapy, and organ transplant. A transplanted organ (kidney) was present in one participant,
which would have been an exclusion criterion. The clinical review committee decided to include the participant in the incidence plots for the primary outcome with (figure 2A)
clinically evaluable population since the transplant was present for more than 10 years. and without non-attributable death considered as a
competing risk factor (figure 2B) show a higher
Table 1: Baseline characteristics of the intention-to-treat and clinically evaluable populations
cumulative incidence in the intravenous group; however,

6 www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9

Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
 Articles

Intention-to-treat population Clinically evaluable population

Oral switch Intravenous Percentage-point Oral switch Intravenous Percentage-point
group (n=108) group (n=105) difference (95% CI) group (n=86) group (n=79) difference (95% CI)
Primary endpoint
SAB-related complication within 14 (13%) 13 (12%) 0·7 (–7·8 to 9·1) 3 (4%) 4 (5%) –2·9 (–9·6 to 3·9)
90 days
Reason primary outcome was met
SAB-related complication 6 (6%) 8 (8%) –2·1 (–9·7 to 5·5) 3 (4%) 4 (5%) –1·6 (–9·0 to 5·8)
Relapsing SAB 3 (3%) 4 (4%) –1·0 (–6·8 to 4·7) 2 (2%) 2 (3%) –0·2 (–5·1 to 4·7)
Deep-seated infection with 5 (5%) 8 (8%) –3·0 (–10·4 to 4·4) 3 (4%) 4 (5%) –1·6 (–9·0 to 5·8)
S aureus
Death attributable to SAB 2 (2%) 0 1·9 (–1·6 to 5·3) 1 (1%) 0 1·2 (–2·3 to 4·6)
Missing outcome data 8 (7%) 5 (5%) 2·7 (–4·7 to 10·0) ∙∙ ∙∙ ∙∙
Attributability of death 3 (3%) 1 (1%) 1·8 (–2·7 to 6·4) ∙∙ ∙∙ ∙∙
non-evaluable
Secondary endpoints
Length of hospital stay from SAB 12 (9–19) 16 (10–19) –2 (–4 to 0); p=0·043* 11 (9–16) 15 (10–18) –2 (–5 to 0); p=0·020*
onset, days
Participants with complications of intravenous administration†
Any complication 9 (9%); 11 17 (17%); 5 –7·9 (–17·6 to 1·9) 6 (7%); 3 13 (17%); 2 –9·5 (–20·5 to 1·5)
Chemical phlebitis 7 9 –2·1 (–10·1 to 5·9) 5 8 –4·3 (–13·8 to 5·2)
Infectious thrombophlebitis or 0 2 –1·9 (–5·5 to 1·7) 0 2 –2·5 (–7·2 to 2·2)
phlebitis
Other‡ 2 6 –3·9 (–9·9 to 2·2) 1 3 –2·6 (–8·6 to 3·4)
Participants with Clostridium 2 (2%); 8 2 (2%); 7 –0·1 (–3·8 to 3·6) 2 (2%); 7 1 (1%); 5 1·1 (–4·0 to 6·1)
difficile infection§
Survival
14 days 98·1% (1·3); 2 100·0% (0); 0 –1·9 (–4·5 to 0·7) 98·8% (1·2); 1 100·0% (0); 0 –1·2 (–3·4 to 1·1)
30 days 94·3% (2·3); 6 96·0% (2·0); 4 –1·7 (–7·6 to 4·2) 98·8% (1·2); 1 98·7% (1·3); 1 0·1 (–3·3 to 3·5)
90 days 83·6% (3·6); 17 89·0% (3·1); 11 –5·4 (–14·8 to 4·0) 92·9% (2·8); 6 94·8% (2·5); 4 –1·9 (–9·3 to 5·5)
Data are n (%), median (IQR), n (%); missing, or Kaplan-Meier estimate (SE); number of deaths, unless specified otherwise. SAB=Staphylococcus aureus bloodstream infection.
*Data are median difference (95% CI). †Complications of intravenous administrations might be caused by study drug or other intravenous medications. ‡In the intention-to-
treat population, other complications of intravenous therapy were haematoma (n=1) and inflammation at a previous peripheral catheter insertion site (n=1) in the oral
switch group and extravasation (n=1), refusal of renewed catheter placement (n=3), and clogged intravenous access (n=2) in the intravenous group. §Missing participants
were those with diarrhoea that did not undergo C difficile testing and were considered negative for statistical analysis.

Table 2: Primary and secondary outcome variables for the intention-to-treat and clinically evaluable populations

this difference was not significant. Details of all
participants with complications related to S aureus
bloodstream infection are shown in appendix 5
(pp 21–22).
The length of hospital stay after the first positive blood
culture was shorter in the oral switch group than in the
intravenous group, with a median stay of 12 days
(IQR 9–19) versus 16 days (10–19; median difference
–2 days, 95% CI –4 to 0; p=0·043; table 2). The shorter
hospital stay in the oral switch group corresponds to the
separation of the curves in the second week, as shown in
the Kaplan-Meier plot of the cumulative incidence of
discharge or death (figure 2C). Orally treated participants
had numerically fewer complications of intravenous
therapy (treatment difference –7·9 percentage points,
95% CI –17·6 to 1·9). The incidence of C difficile infection
was similar in both groups. Survival was numerically
lower in the oral switch group than in the intravenous
group, with a treatment difference of –1·7 percentage
points (95% CI –7·6 to 4·2) at 30 days and –5·4 percentage
points (95% CI –14·8 to 4·0) at 90 days (see figure 2D).
Differences between these secondary endpoints were not
significant.
In subgroup analyses, the treatment difference for the
primary endpoint was similar for female and male
participants (appendix 5 pp 11–12, 27). However, a higher
rate of complications related to S aureus bloodstream
infection occurred in female participants than male
participants (11 [17%] of 65 female participants had a
complication related to S aureus bloodstream infection vs
16 [11%] of 148 male participants). Yet, these differences
were not significant (p=0·26; appendix 5 p 27). The
90-day survival was similar between sexes, with nine
deaths (Kaplan-Meier estimate 85·4%, 95% CI
76·6–94·2) in female participants and 19 deaths (86·6%,
80·9–92·3) in male participants. Demographic variables
were also similar between female and male participants,
except for renal failure, which was present in seven (11%)
of 65 female participants and 42 (28%) of 148 male
participants (appendix 5 pp 28–29).

www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9 7

Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
 Articles

A B
0·10 Oral switch group 0·10
Intravenous group

0·08 0·08
Cumulative incidence

0·06 0·06

0·04 0·04

0·02 0·02
HR 0·74 SHR 0·74
95% CI (0·26–2·13) 95% CI (0·26–2·12)
p=0·58 p=0·57
0 0
0 30 60 90 0 30 60 90
Time to event (days) Time to event (days)
Number at risk
(censored)
Oral switch group 108 94 (10) 87 (16) 82 (21)
Intravenous group 105 93 (6) 86 (13) 82 (15)

C D
1·0 1·0

0·8 0·8
Cumulative incidence

Cumulative survival

0·6 0·6

0·4 0·4

0·2 0·2
HR 1·15 HR 1·54
95% CI (0·88–1·5) 95% CI (0·72–3·28)
p=0·32 p=0·27
0 0
0 30 60 90 0 30 60 90
Time to discharge or death (days) Time to event (days)
Number at risk
(censored)
Oral switch group 108 14 (0) 4 (0) 2 (0) 108 98 (4) 91 (5) 86 (5)
Intravenous group 105 13 (0) 2 (0) 2 (0) 105 97 (5) 90 (6) 88 (6)

Figure 2: Estimates of cumulative incidence from the date of the first positive blood culture for the intention-to-treat population
(A) Cumulative incidence of the composite primary endpoint, complications related to Staphylococcus aureus bloodstream infection, with complications defined as
relapsing S aureus infection, deep-seated S aureus infection, or death attributable to S aureus and participants with deaths not attributable to S aureus bloodstream
infection censored (Kaplan-Meier method). (B) Cumulative incidence of the composite primary endpoint, with mortality not attributable to S aureus bloodstream
infection considered a competing event (post-hoc sensitivity analysis, Fine and Gray model). (C) Cumulative incidence of hospital discharge or death. (D) Cumulative
survival (all-cause mortality, Kaplan-Meier method) with HRs estimated from Cox regression. HR=hazard ratio. SHR=subdistribution hazard ratio.

Regarding other subgroups, analyses of the primary infection than in participants with peripheral venous
outcome were similar regardless of whether the bacterial catheter infection, as well as in participants with a
isolate was meticillin susceptible or resistant, the specific comorbidity than in those with no specific
participants had different foci at baseline, they had a comorbidity (appendix 5 pp 11–12).
Charlson Comorbidity Index less than 3 or 3 or more, The length of hospital stay was longer in the
an echocardiography was performed, or they were 16 participants with MRSA bacteraemia (median 19 days
enrolled in different countries (appendix 5 pp 11–12, 26). [IQR 14–28]) than in the 197 participants with MSSA
Although the numbers of participants in the subgroups (median 15 days [9–18]; p=0·016). The Kaplan-Meier
were small, in the oral switch therapy group, higher estimate for 90-day survival in participants with MRSA
numerical rates of complications related to S aureus was lower (74·5%, SD 11·0) than in participants with
bloodstream infection occurred in participants with MSSA (87·2%, 2·4; p=0·15; appendix 5 p 26). The
central venous catheter infection or skin or soft-tissue incidence of complications of intravenous administration

8 www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9

Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
 Articles

was lower in patients with MRSA (one [6%] of 16 with

Oral switch group Intravenous group p value*
MRSA vs 25 (14%) of 197 with MSSA; p=0·70; appendix 5 (n=107) (n=103)
p 26).
Participants Total Participants Total
Overall, 52 (49%) of 107 participants in the oral switch with events events with events events
group and 42 (41%) of 103 in the intravenous group
Any adverse event 52 (49%) 89 42 (41%) 61 0·27
reported an adverse event (p=0·27); 36 (34%) and
Any serious adverse event 36 (34%) 60 27 (26%) 40 0·29
27 (26%), respectively, were serious adverse events
Drug related† 3 (3%) 3 0 0 ··
(p=0·29; table 3, appendix 5 pp 23–24). The most
Infections and infestations 23 (22%) 27 17 (17%) 18 0·38
common adverse events were infections, with 45 events
Sepsis, septic shock, bacteraemia, and ∙∙ 7 ∙∙ 5 ∙∙
recorded in 40 (19%) participants (table 3, appendix 5 endocarditis
pp 30–32). The rate of serious adverse events per person
Urinary tract infection and cystitis ∙∙ 5 ∙∙ 3 ∙∙
per month was 0·21 (95% CI 0·15–0·29) in the oral
Respiratory tract infection, pneumonia, ∙∙ 4 ·· 2 ∙∙
switch group and 0·14 (0·10–0·20) in the intravenous and bronchitis
group, with a rate ratio of 1·54 (95% CI 0·95–2·50; Wound infection and skin infection ∙∙ 4 ∙∙ 2 ∙∙
p=0·079). Arthritis and osteomyelitis ∙∙ 2 ∙∙ 2 ∙∙
C difficile infection ∙∙ 1 ∙∙ 1 ∙∙
Discussion General disorders and administration site 11 (10%) 13 4 (4%) 4 0·11
In this multicentre, randomised trial in patients with conditions
low-risk S aureus bloodstream infection, an early switch Fever ∙∙ 5 ∙∙ 1 ∙∙
to oral antimicrobial therapy was non-inferior to Oedema ∙∙ 3 ∙∙ 0 ∙∙
continuing standard intravenous therapy regarding the Asthenia and general physical ∙∙ 2 ∙∙ 1 ∙∙
primary endpoint. Participants who received oral deterioration
medication were discharged from hospital earlier than Severe pain ∙∙ 2 ·· 0 ··
participants who received intravenous medication, as Cardiac disorders 9 (8%) 9 5 (5%) 5 0·41
expected from previous trials on oral switch antimicrobial Cardiac and cardiorespiratory arrest ∙∙ 4 ·· 1 ∙∙
therapy in patients with various infections.11,31 We further Cardiac failure ∙∙ 1 ·· 3 ∙∙
found that the rate of complications of intravenous Myocardial ischaemia and myocardial ∙∙ 2 ·· 0 ∙∙
administration was lower in the oral switch group, infarction
although the difference was not significant. We consider Respiratory, thoracic, and mediastinal 4 (4%) 4 10 (10%) 10 0·10
disorders
this outcome to be a true treatment effect rather than due
Pulmonary embolism ∙∙ 1 ∙∙ 2 ∙∙
to chance since with the switch to oral therapy, the
number of intravenously administered drugs that can Dyspnoea ∙∙ 1 ∙∙ 1 ∙∙

potentially cause a complication is reduced. Exacerbated chronic obstructive ∙∙ 1 ∙∙ 1 ∙∙

Survival at 90 days was numerically lower in the oral
switch group than in the intravenous group in both the
intention-to-treat and clinically evaluable analyses
(table 2), albeit not significantly. The two participants
with death attributable to S aureus infection were in the
oral switch group. Both participants had catheter-related
S aureus infection and developed disseminated infection
within 3 weeks after starting study medication.
One participant was discharged and declined readmission
when the general condition deteriorated. This finding
highlights the importance of monitoring patients with
S aureus bloodstream infection closely for developing
complications.
We further observed a higher incidence of adverse
events and serious adverse events in the oral switch
group than in the intravenous group, although the
difference was not significant and the overall incidence
was in the range of previous trials.32,33 A large proportion
of adverse events were infections not caused by S aureus,
which underscores that the participant population is
generally at risk for infections. The numerically lower
survival and the concordant higher incidence of adverse
events in the oral switch group were not significant and,
thus, might be due to chance alone. We cannot exclude
pulmonary disease
Common Terminology Criteria for Adverse Events were recorded from grade 3 and above and listed by System Organ
Class and preferred term. System Organ Class terms are shown when at least 5% of all participants had a serious
adverse event in the respective class. Preferred terms are shown when at least two serious adverse events occurred in
the preferred term. Preferred terms were aggregated as denoted. *Fisher’s exact test compared the number of
participants with at least one event. †Drug-related serious adverse events (classified as certain or probable) were
cardiac failure, Clostridium difficile colitis, and fever. Some participants had more than one adverse event so numbers
might not add up.
Table 3: Adverse events in the safety population occurring within 90 days
that the oral medication affects outcome, although oral
regimens are generally considered to have a favourable
safety profile. Nevertheless, future trials should confirm
equivalence of oral therapy regarding all-cause mortality.
When comparing subgroups of male and female
participants, we found a similar treatment difference for
the primary outcome and similar 90-day survival.
However, the observed higher rate of complications
related to S aureus bloodstream infection in female
participants than male participants might warrant
further studies.
Assigning the low-risk category to patients with
S aureus bloodstream infection can be challenging. In
a randomised controlled trial in patients with

www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9 9

Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
 Articles
uncomplicated staphylococcal bacteraemia, a third of
patients without suspected metastatic infection at
enrolment were later diagnosed with complicated
S aureus bloodstream infection.33 14 (7%) of 213 patients
who met our criteria for low-risk S aureus bloodstream
infection at enrolment had complications related to
S aureus infection (table 2), of which more than half
became apparent during the intervention phase (ie,
during the second week of illness). These early
complications might have been present at baseline and,
unlike in our study, it has been proposed to consider
complications occurring within 5–7 days post-enrolment
as a baseline condition.34 When we excluded early
complications from the primary endpoint in a sensitivity
analysis, we showed non-inferiority (appendix 5
pp 13–14).
The strengths of this study are the pragmatic approach
with a globally available oral medication as the first
choice, the strict definition of low-risk S aureus
bloodstream infection, and the choice of a composite
primary endpoint specific to S aureus bloodstream
infection. The trial has several limitations. It was open
label since intravenous administration of placebo in the
oral switch group would have interfered with the aim of
facilitating early discharge. This study design poses a risk
of bias from investigators when assessing outcomes and
safety endpoints. To mitigate this bias, the clinical review
committee, who were masked to treatment assignment,
assessed the outcomes. In an open-label design, patients
might also introduce bias (eg, by withdrawing informed
consent when not randomised into the desired study
group). However, as two participants from each group
withdrew informed consent, the risk of bias was low.
The large number of patients screened raises the
question of whether low-risk S aureus bloodstream
infection is clinically relevant. However, screening-to-
enrolment ratios above 1:25 are typical for trials on
uncomplicated S aureus bloodstream infection;34
upcoming platform trials might allow more efficient
recruitment.35 A study conducted in one of the
participating centres found that the proportion of
patients with multiple risk factors for complicated or
difficult-to-treat S aureus bloodstream infection increased
over 14 years,36 suggesting that low-risk S aureus
bloodstream infection has decreased in prevalence. The
reduced recruitment rate could not be compensated for
by increasing the capacity of the clinical network and led
to early termination of the trial. To this end, the planned
interim analysis was converted into the final analysis,
which increased the attainable non-inferiority margin
to 10%.
Generally, a combined endpoint is considered less
robust than a single endpoint. However, this design
decision was made consciously to reflect sequelae that
are specific to S aureus bloodstream infection and to
ensure that the required sample size could be handled by
the clinical network.
10 www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9
Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
Due to the relatively low rates of MRSA in participating
countries, we enrolled only 16 participants with MRSA
infection. This raises the question of whether the results
are generalisable to patients with low-risk bloodstream
infection due to MRSA. As in many studies, we found a
lower 90-day survival in participants with MRSA than
MSSA.2 It is important to note that the interpretation of
the primary endpoint did not change when restricting
the analysis to MSSA only.
The results should not be generalised to other oral
medications or to other populations, such as injection
drug users, who were not included in the study due to an
elevated risk for the presence or development of
complicated infection, although these populations might
be perceived as good candidates for oral antimicrobial
therapy. It is important to note that we included more
patients with peripheral venous catheter infection than
with central venous catheter infection, although central
venous catheter infection is usually a more frequent
focus in S aureus bloodstream infection.37 We assume
that this difference was driven by our selection of a low-
risk population, since patients that require central
venous access will have a higher risk of comorbidities
that might lead to exclusion.
Misclassification of low-risk S aureus bloodstream
infection might impede safe implementation of the
findings into clinical practice. A more aggressive
diagnostic approach (eg, involving fluorodeoxyglucose-
PET or CT),38 a more stringent evaluation of the need for
echocardiography (eg, the VIRSTA study score),39 or
novel biomarkers40 might identify and predict deep-
seated infections more accurately. Nevertheless, careful
assessment of the individual, preferably by consulting an
infectious diseases expert, is necessary to correctly assign
a patient to the low-risk group, and close monitoring
for early detection of complications and potential
prolongation of antimicrobial therapy are advised.
In summary, we found that an early oral switch was
non-inferior to standard intravenous antimicrobial
therapy in patients with low-risk S aureus bloodstream
infection. This study supports an early switch to oral
antimicrobial therapy in patients with low-risk S aureus
bloodstream infection provided a rigorous clinical
assessment and close monitoring for complications are
done. Future studies should address other subgroups of
patients with S aureus bloodstream infection and confirm
the benefits of early oral switch in a detailed medico-
economic analysis.
SABATO study group
Adoración Valiente, Marina de Cueto (Hospital Universitario Virgen
Macarena, Seville, Spain), Ángel Rodríguez, José Molina (Hospital
Universitario Virgen del Rocío, Seville, Spain), Julia Fischer (University
of Münster, Germany), Gregor Paul (University of Cologne, Cologne,
Germany) Reinhild Prinz-Langenohl (Clinical Trials Centre Cologne,
Cologne, Germany), Gabriele Peyerl-Hoffmann, Daniel Hornuß
(University Medical Centre Freiburg, Freiburg, Germany), Sébastien
Gallien, Vincent Fihman (Henri-Mondor University Hospital, AP-HP,
Creteil, France), Marion Lacasse, Francois Coustillères (Centre
Hospitalier Régional Universitaire de Tours, Tours, France),

Christian Becker (Helios Klinikum Krefeld, Krefeld, Germany),
André Fuchs (University Hospital of Augsburg, Augsburg, Germany),
Laura Morata (Hospital Clínic, Barcelona, Spain), Sebastian Weis (Jena
University Hospital—Friedrich Schiller University Jena, Jena, Germany),
Diane Ponscarme (Saint-Louis Hospital, Paris, France),
Lydie Khatchatourian (Centre Hospitalier de Cornouaille, Quimper,
France), Elisabeth Rouveix (Ambroise Paré University Hospital,
Boulogne-Billancourt, France), Dominique Merrien (Centre Hospitalier
Départmentale Vendée, La Roche-sur-Yon, France), Raphaël Lecomte
(Centre Hospitalier Universitaire de Nantes, Nantes, France),
Jacobien Veenemans (Amphia Hospital Breda, CK Breda, Netherlands),
Helem H Vilchez (Fundació Institut d’Investigació Sanitària Illes
Balears, Palma de Mallorca, Spain), Violaine Tolsma (Centre Hospitalier
d’Annecy Genevois, France), Johanna Kessel (University Hospital
Frankfurt, Frankfurt, Germany), Marc J M Bonten (University Medical
Center Utrecht, Utrecht, Netherlands), Jan Rupp (University Hospital of
Schleswig Holstein, Luebeck, Germany), Laurent Hocqueloux (Centre
Hospitalier Régional, Orléans, France), Frederic Lucht (Centre
Hospitalier Universitaire de Saint-Étienne, France), Jean-Paul Stahl
(Centre Hospitalier Universitaire Grenoble, Grenoble, France),
Anne L M Vlek (Diakonessenhuis, Utrecht, Netherlands).
Contributors
AJK was the coordinating principal investigator and sponsor’s
representative. GF served as the clinical trial lead until July, 2016.
AJK, HSe, WVK, SRi, and MH conceived and designed the study.
AJK, AR, and MN managed the trial. AJK, LEL-C, JR-B, JMC, MDN, GF,
NJ, WVK, SRi, RL, LC, LB, AL, KK, CRM, AS, SH, BF, ML, J-PT, AD,
TG, DB, TW, SRe, JK, MLM, EF, HSt, VV, and PT were site investigators
involved in participant recruitment and data collection. MH, AA, and
AJK developed the statistical analysis plan. AJK, AR, AA, and
MH curated the data. MH and AA had access to the data in the study
and conducted the formal statistical analysis. AJK, MH, and AA verified
the data with the clinical review committee and the monitoring team.
AJK, MH, AA, and HSe wrote the first draft of the manuscript.
All authors had full access to all the data in the study and had final
responsibility for the decision to submit for publication.
Declaration of interests
AJK received funding for this study from the Deutsche
Forschungsgemeinschaft and is Chairperson of the German Sepsis
Society. HSe received grants or research support from the
Bundesministerium für Bildung und Forschung (BMBF) Germany and
the German Center for Infection Research, and has been a consultant
for Debiopharm, Gilead, MSD, and Shionogi. AS has received grants
from Gilead Sciences (IN-ES-540–6089) and Pfizer, consulting fees and
lecture honoraria from Pfizer, MSD, Angelini, Shionogi, Gilead, and
Menarini, and travel support from Pfizer. JR-B has received grants or
research support from the Spanish Network for Research in Infectious
Diseases (REIPI RD16/0016/00001) and Centro de Investigación
Biomédica en Red de Enfermedades Infecciosas (CIBERINFEC,
CB21/13/00012), Instituto de Salud Carlos III, Spanish Ministry of
Science and Innovation, which are co-financed by the European
Development Regional Fund. SH received honoraria and travel support
from Shionogi, Pfizer, Infectopharm, and AdvanzPharma. NJ reports
receiving honoraria for lectures from AbbVie, Bayer, Infectopharm, and
Medacta, travel support from Gilead, Pfizer, and Correvio, participation
in paid advisory board meetings for MSD, and membership in the
steering committee of the German Society of Internal Medicine.
SRi received honoraria for lectures from Falk Foundation, Pfizer,
bioMérieux, Akademie für Infektionsmedizin, Med Update,
streamedup!, and Deutscher Apotheker-Verlag. TW reports receiving
honoraria for presentations from AstraZeneca, Advanz, MSD, Pfizer,
and Shionogi, grants or research support from BMBF Germany, and
travel support from Pfizer. All other authors declare no competing
interests.
Data sharing
Individual participants’ data (including the data dictionary) that
underlie the results reported in this Article will be shared on request
after de-identification. The study protocol, statistical analysis plan, and
informed consent form will also be shared. Data will be available
12 months after publication and ending 36 months following
www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9
Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
Articles
publication. Data will be shared for any analysis with researchers who
submit a methodologically sound proposal, which is approved by the
ethics committee at the institution of the corresponding author.
Requests should be directed to the corresponding author of the Article.
To gain access, a data sharing agreement will need to be signed. Data
will then be available from the University data warehouse.
Acknowledgments
AJK received funding for this trial from Deutsche
Forschungsgemeinschaft (217917502). We thank the participants and staff
from all sites who contributed to the trial (appendix 5 pp 3–4). We thank
Vance G Fowler, Stephan Harbarth, and Guy Thwaites for the expert
opinion in the trial steering committee; Vincent LeMoing,
Miquel Pujol Rojo, and Estée Török for participating in the clinical review
committee; and Walter E Haefeli, Alexandra Heininger, and
Geraldine Rauch for participating in the data monitoring committee.
This trial was supported by Clinical Trials Centre Cologne (Medical
Faculty, University of Cologne) with regard to project management,
database development, data management, monitoring, and safety
management.
References
1 Kern WV, Rieg S. Burden of bacterial bloodstream
infection—a brief update on epidemiology and significance of
multidrug-resistant pathogens. Clin Microbiol Infect 2020;
26: 151–57.
2 Bai AD, Lo CKL, Komorowski AS, et al. Staphylococcus aureus
bacteraemia mortality: a systematic review and meta-analysis.
Clin Microbiol Infect 2022; 28: 1076–84.
3 Kaasch AJ, Barlow G, Edgeworth JD, et al. Staphylococcus aureus
bloodstream infection: a pooled analysis of five prospective,
observational studies. J Infect 2014; 68: 242–51.
4 Kaasch AJ, Kern WV, Joost I, Hellmich M, Seifert H, Rieg S. Effect
of clinically uninfected orthopedic implants and pacemakers/
AICDs in low-risk Staphylococcus aureus bloodstream infection on
crude mortality rate: a post hoc analysis of a large cohort study.
Open Forum Infect Dis 2019; 6: ofz170.
5 Bupha-Intr O, Blackmore T, Bloomfield M. Efficacy of early oral
switch with β-lactams for low-risk Staphylococcus aureus bacteremia.
Antimicrob Agents Chemother 2020; 64: e02345–19.
6 Willekens R, Puig-Asensio M, Ruiz-Camps I, et al. Early oral switch
to linezolid for low-risk patients with Staphylococcus aureus
bloodstream infections: a propensity-matched cohort study.
Clin Infect Dis 2019; 69: 381–87.
7 Liu C, Bayer A, Cosgrove SE, et al. Clinical practice guidelines by
the Infectious Diseases Society of America for the treatment of
methicillin-resistant Staphylococcus aureus infections in adults and
children. Clin Infect Dis 2011; 52: e18–55.
8 Gudiol F, Aguado JM, Almirante B, et al. Executive summary of the
diagnosis and treatment of bacteremia and endocarditis due to
Staphylococcus aureus. A clinical guideline from the Spanish Society
of Clinical Microbiology and Infectious Diseases (SEIMC).
Enferm Infecc Microbiol Clin 2015; 33: 626–32.
9 Brown NM, Goodman AL, Horner C, Jenkins A, Brown EM.
Treatment of methicillin-resistant Staphylococcus aureus (MRSA):
updated guidelines from the UK. JAC Antimicrob Resist 2021;
3: dlaa114.
10 Dagher M, Fowler VG Jr, Wright PW, Staub MB. A narrative review
of early oral stepdown therapy for the treatment of uncomplicated
Staphylococcus aureus bacteremia: yay or nay? Open Forum Infect Dis
2020; 7: ofaa151.
11 Iversen K, Ihlemann N, Gill SU, et al. Partial oral versus
intravenous antibiotic treatment of endocarditis. N Engl J Med 2019;
380: 415–24.
12 Li H-K, Rombach I, Zambellas R, et al. Oral versus intravenous
antibiotics for bone and joint infection. N Engl J Med 2019;
380: 425–36.
13 Jorgensen SCJ, Lagnf AM, Bhatia S, Shamim M-D, Rybak MJ.
Sequential intravenous-to-oral outpatient antibiotic therapy for
MRSA bacteraemia: one step closer. J Antimicrob Chemother 2019;
74: 489–98.
14 Kouijzer IJE, van Leerdam EJ, Gompelman M, et al. Intravenous to
oral switch in complicated Staphylococcus aureus bacteremia without
endovascular infection: a retrospective single-center cohort study.
Clin Infect Dis 2021; 73: 895–98.
11
 Articles
15 Mun SJ, Kim S-H, Huh K, et al. Oral step-down therapy in patients
with uncomplicated Staphylococcus aureus primary bacteremia and
catheter-related bloodstream infections. J Chemother 2022;
34: 319–25.
16 Yeager SD, Oliver JE, Shorman MA, Wright LR, Veve MP.
Comparison of linezolid step-down therapy to standard parenteral
therapy in methicillin-resistant Staphylococcus aureus bloodstream
infections. Int J Antimicrob Agents 2021; 57: 106329.
17 Pérez-Rodríguez MT, Sousa A, Moreno-Flores A, et al. The benefits
and safety of oral sequential antibiotic therapy in non-complicated
and complicated Staphylococcus aureus bacteremia. Int J Infect Dis
2021; 102: 554–60.
18 Itoh N, Hadano Y, Saito S, Myokai M, Nakamura Y, Kurai H.
Intravenous to oral switch therapy in cancer patients with catheter-
related bloodstream infection due to methicillin-sensitive
Staphylococcus aureus: a single-center retrospective observational
study. PLoS One 2018; 13: e0207413.
19 Diego-Yagüe I, Mora-Vargas A, Vázquez-Comendador JM, et al.
Sequential oral antibiotic in uncomplicated Staphylococcus aureus
bacteraemia: a propensity-matched cohort analysis.
Clin Microbiol Infect 2023; 29: 744–50.
20 Liu C, Strnad L, Beekmann SE, Polgreen PM, Chambers HF.
Clinical practice variation among adult infectious disease
physicians in the management of Staphylococcus aureus bacteremia.
Clin Infect Dis 2019; 69: 530–33.
21 Hospenthal DR, Waters CD, Beekmann SE, Polgreen PM. Practice
patterns of infectious diseases physicians in transitioning from
intravenous to oral therapy in patients with bacteremia.
Open Forum Infect Dis 2019; 7: ofz386.
22 Kaasch AJ, Fätkenheuer G, Prinz-Langenohl R, et al. Early oral
switch therapy in low-risk Staphylococcus aureus bloodstream
infection (SABATO): study protocol for a randomized controlled
trial. Trials 2015; 16: 450.
23 Kaasch AJ, Rommerskirchen A, Hellmich M, et al. Protocol update
for the SABATO trial: a randomized controlled trial to assess early
oral switch therapy in low-risk Staphylococcus aureus bloodstream
infection. Trials 2020; 21: 175.
24 Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ
failure and guidelines for the use of innovative therapies in sepsis.
Chest 1992; 101: 1644–55.
25 US Department of Health and Human Services, National Institutes
of Health, National Cancer Institute. Common terminology criteria
for adverse events Version 4.0. Bethesda, MD: National Cancer
Institute, 2009.
26 Zhao G. Tests of non-null hypothesis on proportions for stratified
data. Stat Med 2008; 27: 1429–46.
27 European Medicines Agency. Guideline on the evaluation of
medicinal products indicated for treatment of bacterial infections—
revision 2. Dec 15, 2011. https://www.ema.europa.eu/en/
documents/scientific-guideline/guideline-evaluation-medicinal-
products-indicated-treatment-bacterial-infections-revision-2_en.pdf
(accessed Dec 21, 2023).
12 www.thelancet.com/infection Published online January 17, 2024 https://doi.org/10.1016/S1473-3099(23)00756-9
Descargado para Anonymous User (n/a) en Austral University Faculty of Biomedical Sciences de ClinicalKey.es por Elsevier en enero 30, 2024. Para
uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2024. Elsevier Inc. Todos los derechos reservados.
28 European Medicines Agency. Points to consider on switching
between superiority and non-inferiority. July 27, 2000. https://www.
ema.europa.eu/en/documents/scientific-guideline/points-consider-
switching-between-superiority-and-non-inferiority_en.pdf (accessed
Dec 21, 2023).
29 European Medicines Agency. Guideline on the evaluation of
medicinal products indicated for treatment of bacterial infections—
revision 3. May 19, 2022. https://www.ema.europa.eu/en/
documents/scientific-guideline/guideline-evaluation-medicinal-
products-indicated-treatment-bacterial-infections-revision-3_en.pdf
(accessed Dec 21, 2023).
30 Fine JP, Gray RJ. A proportional hazards model for the
subdistribution of a competing risk. J Am Stat Assoc 1999;
94: 496–509.
31 Schrenzel J, Harbarth S, Schockmel G, et al. A randomized clinical
trial to compare fleroxacin–rifampicin with flucloxacillin or
vancomycin for the treatment of staphylococcal infection.
Clin Infect Dis 2004; 39: 1285–92.
32 Thwaites GE, Scarborough M, Szubert A, et al. Adjunctive
rifampicin for Staphylococcus aureus bacteraemia (ARREST):
a multicentre, randomised, double-blind, placebo-controlled trial.
Lancet 2018; 391: 668–78.
33 Holland TL, Raad I, Boucher HW, et al. Effect of algorithm-based
therapy vs usual care on clinical success and serious adverse events
in patients with staphylococcal bacteremia: a randomized clinical
trial. JAMA 2018; 320: 1249–58.
34 Holland TL, Chambers HF, Boucher HW, et al. Considerations for
clinical trials of Staphylococcus aureus bloodstream infection in
adults. Clin Infect Dis 2019; 68: 865–72.
35 Tong SYC, Mora J, Bowen AC, et al. The Staphylococcus aureus
Network Adaptive Platform trial protocol: new tools for an old foe.
Clin Infect Dis 2022; 75: 2027–34.
36 Mathé P, Göpel S, Hornuss D, et al. Increasing numbers and
complexity of Staphylococcus aureus bloodstream infection—14 years
of prospective evaluation at a German tertiary care centre with
multi-centre validation of findings. Clin Microbiol Infect 2023;
29: 1197.e9–15.
37 Nambiar K, Seifert H, Rieg S, et al. Survival following
Staphylococcus aureus bloodstream infection: a prospective
multinational cohort study assessing the impact of place of care.
J Infect 2018; 77: 516–25.
38 Ghanem-Zoubi N, Kagna O, Abu-Elhija J, et al. Integration of
FDG-PET/CT in the diagnostic workup for Staphylococcus aureus
bacteremia: a prospective interventional matched-cohort study.
Clin Infect Dis 2021; 73: e3859–66.
39 Tubiana S, Duval X, Alla F, et al. The VIRSTA score, a prediction
score to estimate risk of infective endocarditis and determine
priority for echocardiography in patients with Staphylococcus aureus
bacteremia. J Infect 2016; 72: 544–53.
40 Cao Y, Guimaraes AO, Peck MC, et al. Risk stratification biomarkers
for Staphylococcus aureus bacteraemia. Clin Transl Immunology 2020;
9: e1110.