Is there a role for anticoagulation with dabigatran in S. aureus bacteremia? Protocol for the adjunctive treatment domain of the Staphylococcus aureus Network Adaptive Platform (SNAP) randomised controlled trial Authors: Emily G McDonald, Matthew P Cheng, Joshua S Davis, Anna L Goodman, Patrick R Lawler, Julie Marsh, Dominik Mertz, Mical Paul, Jesús Rodriguez-Baño, Deborah M Siegal, Steven YC Tong, Genevieve Walls, Todd C Lee, Marc Bonten Published: 2025 ABSTRACT AbstractIntroductionMany patients receive oral anticoagulation for reduced stroke risk in atrial fibrillation or as treatment or prevention of venous thromboembolism. Oral factor Xa inhibitors (oral FXaI, eg, apixaban, edoxaban or rivaroxaban) are commonly prescribed for this indication. Dabigatran, an oral direct thrombin inhibitor, is similarly approved. In vitro and animal model evidence suggests that dabigatran also has direct effects on Staphylococcus aureus virulence and infection. Observational data have shown that dabigatran users are less likely to develop S. aureus bacteremia (SAB), and a small randomised controlled trial showed that dabigatran has anti-S. aureus effects when compared with low molecular weight heparins during bloodstream infection. We seek to answer whether dabigatran is superior to the oral FXaIs in achieving better SAB outcomes among patients who independently require oral anticoagulation. We report the intervention-specific protocol, embedded in an adaptive platform trial.Methods and analysisThe S. aureus Network Adaptive Platform (SNAP) trial [NCT05137119] is a pragmatic, randomised, multicentre adaptive platform trial that compares different SAB therapies for 90-day mortality rates. For this intervention (‘Dabi-SNAP’), patients receiving therapy with an oral FXaI will be randomised to continue as usual or to change to dabigatran as of the next scheduled dose. All subjects will receive standard of care antibiotics and/or antibiotics allocated through other active domains in the platform. As the choice of anticoagulant may not demonstrate large differences in mortality, a ranked composite of death and adverse outcomes (Desirability of Outcome Ranking, or DOOR) was chosen as the primary outcome.Ethics and disseminationThe study is conditionally approved by the research ethics board of the McGill University Health Centre: identifier 2025-10900. Trial results will be published open access in a peer-reviewed journal and presented at a global infectious disease conference. The trial is registered at clinicaltrials.gov with the identifier NCT06650501.Trial registration numberNCT0665050. STRENGTHS AND LIMITATIONS OF THIS STUDY The study uses a randomised clinical trial design to evaluate anticoagulation in patients with Staphylococcus aureus bacteraemia, minimising bias and confounding. The study uses a randomised clinical trial design to evaluate anticoagulation in patients with Staphylococcus aureus bacteraemia, minimising bias and confounding. A current-user design was adopted, recruiting participants who had already tolerated anticoagulation, which enhances internal validity and participant safety. A current-user design was adopted, recruiting participants who had already tolerated anticoagulation, which enhances internal validity and participant safety. The primary outcome is based on the Desirability of Outcome Ranking (DOOR) framework, allowing a comprehensive assessment of benefit and harm across multiple clinical domains. The primary outcome is based on the Desirability of Outcome Ranking (DOOR) framework, allowing a comprehensive assessment of benefit and harm across multiple clinical domains. The exclusion of new users of anticoagulants may limit the generalisability of the findings to patients initiating therapy. The exclusion of new users of anticoagulants may limit the generalisability of the findings to patients initiating therapy. Introduction Staphylococcus aureus is a leading cause of bacterial-related mortality worldwide1; adult patients with S. aureus bacteremia (SAB) have a 3-month mortality rate in excess of 26%.2 Until recently, there have been a limited number of randomised controlled trials (RCTs) for SAB.3 This changed with the S. aureus Network Adaptive Platform (SNAP) [NCT 05137119], which has now recruited more patients with SAB than all other RCTs combined.4 Within SNAP, we are comparing a variety of antibiotic strategies for the management of different types of SAB.46 Although antithrombotic adjunctive therapy has proved effective for other infections (eg, COVID-19),7 and has elicited improvements in patient-relevant outcomes (eg, a reduced need for Intensive Care Unit (ICU) level organ support and decreased mortality),8 no RCT has previously studied antithrombotic therapies as adjunctive treatments to improve patient outcomes for patients with SAB. Indeed, adult patients with SAB are at high risk of complications as the population tends to be older and more medically complex. Such patients often have pre-existing conditions which independently serve as an indication for antithrombotic therapy. However, one concern in studying the initiation of antithrombotic therapy in this disease is the risk of bleeding complications. Whereas patients who are already on antithrombotic therapy for pre-existing conditions usually continue their therapy in the presence of SAB unless a new contraindication arises. By leveraging a population of pre-existing medication users with evidence-based indications for treatment, we can test the hypothesis that certain agents, by virtue of specific anti-S. aureus activity, might be beneficial, while simultaneously avoiding an unjustified bleeding risk from initiating new anticoagulation starts. Dabigatran etexilate is an oral direct thrombin inhibitor (Factor IIa) that is licensed in many countries to reduce thrombotic risk in non-valvular atrial fibrillation and to treat and prevent venous thromboembolism (deep vein thrombosis (DVT) or pulmonary embolism (PE)).9 There is evidence that dabigatran interferes with S. aureus virulence and pathogenesis through multiple mechanisms: notably, dabigatran inhibits coagulase10 and staphylothrombin-mediated fibrin deposition.11 In mice, administration of dabigatran reduced S. aureus abscess size11; in experimental S. aureus infective endocarditis in rats, dabigatran reduced the vegetation size, bacterial load and inflammation.12 In humans, a small proof of concept RCT (n=94) showed that the bacterial enzyme staphylothrombin was effectively inhibited at routinely used doses of dabigatran13 when compared with low molecular weight heparins. Finally, a population-level study in Denmark reported that patients on dabigatran had a lower incidence of invasive S. aureus disease relative to those receiving oral factor Xa inhibitors (FXaI; eg, apixaban, edoxaban and rivaroxaban).14 A similar finding was presented in abstract form for veterans with atrial fibrillation.15 Dabigatran and the oral FXaIs are collectively referred to as direct oral anticoagulants (DOACs). These molecules share the same treatment indications; however, in contrast to dabigatran, the oral FXaIs do not directly interfere with bacterial virulence factors. There have been no large published RCTs directly comparing the FXaIs with non-FXaI DOACs and, therefore, guidelines16 17 generally do not recommend preferential use of one DOAC over another within currently approved indications. Use in clinical practice is therefore influenced by drug characteristics (eg, frequency of administration, renal clearance, drug interactions), patient characteristics (eg, age, allergies, weight, renal function, liver function), patient/provider preference and cost/reimbursement. Most patients can transition safely between DOACs.18 Indeed, switching DOACs can happen by routine chance if a hospital formulary does not carry all options. In this context, we aim to conduct a superiority, parallel group, 1:1 randomisation, clinical trial nested within the SNAP trial,4 known as Dabi-SNAP, which evaluates the potential for dabigatran to improve outcomes in patients with SAB who otherwise require oral anticoagulation. Methods and analysis Setting Acute care hospitalisation. Population Adult patients who meet the entry criteria for SNAP4 and who have an existing approved indication for oral FXaI therapy. Randomisation Eligible adult patients will be randomised in a 1:1 ratio when their eligibility for the intervention is confirmed. The treatment assignments are generated using permuted blocks, with block sizes of 4, 6 and 8, and stratified by the indication for the oral FXaI (atrial fibrillation vs venous thromboembolism/both) and by any history of upper gastrointestinal bleeding (yes vs no/unknown). Treatment assignment is performed centrally in the study database. In the event a patient requires an antithrombotic drug to be held prior to randomisation (for example, due to a procedure or acute kidney injury), the patient may remain eligible for up to 5 days from the index blood culture, provided the antithrombotic therapy is resumed prior to randomisation. Blinding and unblinding This is an open label study for reasons of safety. Patients and their treating teams need to know which anticoagulants they are taking to ensure proper medication reconciliation is performed, to mitigate drug-drug interactions and to be able to use specific antidotes if there is life-threatening bleeding. Allocation will be concealed until the eligible patient is starting the study drug. On a study-wide basis, investigators, site and study personnel will remain blinded to pooled domain outcomes and summaries until the Data Safety and Monitoring Committee (DSMC) has recommended terminating the study or domain for superiority, signal of harm or if the maximum platform recruitment target is met. Interventions Patients will be randomised to continue their existing oral FXaI as prescribed or to change to dabigatran 150 mg PO (per os) two times per day. In regions where a 110 mg dose of dabigatran is marketed, this dose may be used in patients over age 80 years, or for those 75 years or older with 1 or more risk factors for bleeding (see online supplemental file). Patients may simultaneously be enrolled into any other domain of the platform for which they are eligible. Eligibility criteria/inclusion/exclusion criteria Adult patients who have consented and been enrolled into the platform,4 and who are taking (or will imminently start) an oral FXaI are eligible for inclusion. Exclusion criteria are listed in table 1. Table 1Inclusion and exclusion criteriaInclusion criteriaExclusion criteriaEnrolled in SNAPTaking (or starting imminently) an oral FXaI (apixaban, edoxaban or rivaroxaban) for: stroke prevention in atrial fibrillation, treatment or secondary prevention of deep venous thrombosis or pulmonary embolism, prevention of venous thromboembolism in patients who have undergone elective total hip or total knee replacement surgery provided there are 30 or more days of planned treatment remaining at the time of enrolment.Patients will be excluded from this domain if they have any of the following at the time of eligibility assessment:Active bleeding (as determined by site investigator after discussion with treating team)*Anticipated major cardiac surgery, neurosurgery or spine surgery within the next 3 daysKnown pregnancy (with testing available for women with childbearing potential)Known use of dabigatran within the last monthAllergy to dabigatranConcomitant use of amiodarone, ketoconazole, rifampin, verapamil, clopidogrel, prasugrel or ticagrelorEstimated glomerular filtration rate below 30 mL/min*Off label use of the oral FXaI (eg, metallic mechanical heart valve, left ventricular thrombus, antiphospholipid antibody syndrome)*Patient may remain eligible for up to 120 hours from platform entry if condition is resolved and antithrombotic therapy is resumed.FXaI, factor Xa inhibitor; SNAP, Staphylococcus aureus Network Adaptive Platform. Inclusion and exclusion criteria Active bleeding (as determined by site investigator after discussion with treating team)* Active bleeding (as determined by site investigator after discussion with treating team)* Anticipated major cardiac surgery, neurosurgery or spine surgery within the next 3 days Anticipated major cardiac surgery, neurosurgery or spine surgery within the next 3 days Known pregnancy (with testing available for women with childbearing potential) Known pregnancy (with testing available for women with childbearing potential) Known use of dabigatran within the last month Known use of dabigatran within the last month Allergy to dabigatran Allergy to dabigatran Concomitant use of amiodarone, ketoconazole, rifampin, verapamil, clopidogrel, prasugrel or ticagrelor Concomitant use of amiodarone, ketoconazole, rifampin, verapamil, clopidogrel, prasugrel or ticagrelor Estimated glomerular filtration rate below 30 mL/min* Estimated glomerular filtration rate below 30 mL/min* Off label use of the oral FXaI (eg, metallic mechanical heart valve, left ventricular thrombus, antiphospholipid antibody syndrome) Off label use of the oral FXaI (eg, metallic mechanical heart valve, left ventricular thrombus, antiphospholipid antibody syndrome) Patient may remain eligible for up to 120 hours from platform entry if condition is resolved and antithrombotic therapy is resumed. FXaI, factor Xa inhibitor; SNAP, Staphylococcus aureus Network Adaptive Platform. Patients receiving dabigatran will be excluded for two main reasons. First, if dabigatran is beneficial, we would be randomising patients to a potentially inferior agent in the setting of SAB (potential harm). A trial which concludes “don’t change away from dabigatran because of harm” provides a different conclusion than a trial which concludes “change to dabigatran for benefit”. Second, if dabigatran impacts the life cycle and pathogenesis of S. aureus, patients who develop invasive infection while receiving it may have organisms which have already adapted to the drug. This is akin to a patient developing an infection while already on therapeutic doses of an antibiotic—there is a risk for “resistance”. The patients who are already users may have a differential response to therapy because of this. Relatedly, if dabigatran prevents SAB, the subgroup will accrue at a slower rate. Data collection and participant timeline Along with data collected in the core SNAP protocol,4 and for additional domains where consent has been obtained, patients will be followed regularly (every 2–3 days) in hospital and fortnightly post discharge (virtual or in person as required to promote retention in the trial) until platform day 90 for the development of bleeding and thrombotic events (table 2). Participants who discontinue the study drug or deviate from planned study protocols will continue to be followed until day 90 for the primary and secondary outcomes. Table 2Participant timelineEventTimelineS. aureus bacteremia identifiedDays −3 to 0EnrolmentDay 0In-hospital follow-upEvery 2–3 daysPost discharge follow-upEvery 2 weeksPrimary and secondary outcomesDay 90S. aureus, Staphylococcus aureus. Participant timeline S. aureus, Staphylococcus aureus. Outcomes This trial only involves a subpopulation of patients with SAB, and as such, large differences in the SNAP core protocol primary endpoint, mortality at day 90, are not anticipated. Consequently, the primary outcome for this nested RCT is the ranked Desirability of Outcome Ranking (DOOR) at day 90 (figure 1) similar to one previously developed for SAB.19 Lower ranks for the DOOR indicate better SAB outcomes. Desirability of Ordinal Outcome Ranking (DOOR). In Step 1 of conducting the DOOR scores, clinical success is defined as resolution of clinical signs and symptoms of SAB, such that no additional antibiotic therapy is required or anticipated for its treatment. Clinical failure is, therefore, the absence of success. Infectious complications refer to the new (ie, no indication of presence at platform entry) development of conditions between platform entry and day 90. Such conditions include: endocarditis; new evidence of metastatic foci (eg, osteomyelitis, deep abscess); relapse (the isolation of a concordant SAB from a blood culture drawn after a patient had become negative with therapy); readmission for subsequent care of SAB, or need for unplanned source control procedures. Serious adverse events as graded 4+ by the Common Terminology Criteria for Adverse Events (CTCAE)28 (which are probably or definitely related to study drug20), and adverse events refer to a non-serious event leading to study drug discontinuation by the treating team. Study drug for these assessments only refers to anticoagulants (FXaI or dabigatran) and does not consider other domain-specific interventions. Placing the study drug on hold for up to 5 days for clinical reasons (eg, for a procedure) is permitted and will not be classified as a discontinuation. In Step 2, surviving participants at day 90 with the same DOOR score (ie, ties) are ranked according to the inpatient length of stay for the index hospitalisation, ensuring that shorter hospital stay is better (ie, lower rank) but not at the expense of clinical outcomes. Key secondary core endpoints from SNAP and the secondary endpoints specific to this trial are shown in box 1. Secondary endpoints Individual DOOR components at day 90 Clinical failure. Clinical failure. Infectious complication. Infectious complication. Serious adverse event or adverse event leading to study drug discontinuation. Serious adverse event or adverse event leading to study drug discontinuation. All cause mortality. All cause mortality. Other endpoints All cause mortality at days 14, 28 and 42. All cause mortality at days 14, 28 and 42. Index hospitalisation length of stay among survivors. Index hospitalisation length of stay among survivors. Clinically relevant bleeding within 90 days. Clinically relevant bleeding within 90 days. Defined as one or more of the following: fatal bleeding; symptomatic bleeding in a critical area or organ (including haemorrhagic stroke); bleeding that causes a fall in haemoglobin level of ≥20 g/L; or bleeding that requires a transfusion of 2 or more units of whole blood or red cells. Defined as one or more of the following: fatal bleeding; symptomatic bleeding in a critical area or organ (including haemorrhagic stroke); bleeding that causes a fall in haemoglobin level of ≥20 g/L; or bleeding that requires a transfusion of 2 or more units of whole blood or red cells. Clinically relevant venous thromboembolic events within 90 days. Clinically relevant venous thromboembolic events within 90 days. Defined as an acute objectively confirmed deep vein thrombosis (upper extremity, lower extremity, other such as portal vein, cerebral vein, splanchnic vein) and/or pulmonary embolism, which is symptomatic and/or necessitates specific treatment. Below-knee DVT is not included. Pulmonary embolism needs to be segmental or proximal. Defined as an acute objectively confirmed deep vein thrombosis (upper extremity, lower extremity, other such as portal vein, cerebral vein, splanchnic vein) and/or pulmonary embolism, which is symptomatic and/or necessitates specific treatment. Below-knee DVT is not included. Pulmonary embolism needs to be segmental or proximal. Clinically relevant stroke within 90 days. Clinically relevant stroke within 90 days. Defined as the acute onset of focal neurological dysfunction caused by brain, spinal cord or retinal vascular injury because of infarction (ischaemia). Ideally, at least one of the following should be present to confirm the diagnosis of stroke: confirmation by neurology or neurosurgical specialist, brain imaging (eg, CT scan, MRI scan or cerebral vessel angiography) compatible with acute ischaemia). If the acute focal signs represent a worsening of a previous deficit, these signs must persist for more than 24 hours and be accompanied by an appropriate new MRI or CT scan finding. In the absence of neuroimaging, a staff neurologist consult which makes the diagnosis of stroke will be considered. Defined as the acute onset of focal neurological dysfunction caused by brain, spinal cord or retinal vascular injury because of infarction (ischaemia). Ideally, at least one of the following should be present to confirm the diagnosis of stroke: confirmation by neurology or neurosurgical specialist, brain imaging (eg, CT scan, MRI scan or cerebral vessel angiography) compatible with acute ischaemia). If the acute focal signs represent a worsening of a previous deficit, these signs must persist for more than 24 hours and be accompanied by an appropriate new MRI or CT scan finding. In the absence of neuroimaging, a staff neurologist consult which makes the diagnosis of stroke will be considered. Clinically relevant acute myocardial infarction within 90 days. Clinically relevant acute myocardial infarction within 90 days. Patients with S. aureus bacteremia often have substantial physiological stresses which can be associated with rises in cardiac troponin (demand ischaemia). For the purposes of this outcome, acute myocardial infarction is captured as a type 1 myocardial infarction: detection of rise and/or fall of cardiac biomarkers with at least one value above the 99th percentile of the upper reference limit together with evidence of myocardial ischaemia. For the purposes of this pragmatic trial, such a myocardial infarction will be inferred from the opinions of the staff cardiologist, intensivist or general internal medicine specialist. Patients with S. aureus bacteremia often have substantial physiological stresses which can be associated with rises in cardiac troponin (demand ischaemia). For the purposes of this outcome, acute myocardial infarction is captured as a type 1 myocardial infarction: detection of rise and/or fall of cardiac biomarkers with at least one value above the 99th percentile of the upper reference limit together with evidence of myocardial ischaemia. For the purposes of this pragmatic trial, such a myocardial infarction will be inferred from the opinions of the staff cardiologist, intensivist or general internal medicine specialist. Clinically relevant arterial thromboembolic event within 90 days. Clinically relevant arterial thromboembolic event within 90 days. Defined as a clinical history compatible with sudden worsening of end organ or limb perfusion and confirmation by imaging (eg, CT angiography, arterial Doppler) or need for urgent surgery or thrombolysis. Defined as a clinical history compatible with sudden worsening of end organ or limb perfusion and confirmation by imaging (eg, CT angiography, arterial Doppler) or need for urgent surgery or thrombolysis. DOOR, Desirability of Outcome Ranking; DVT, deep vein thrombosis. Sample size and recruitment Sample size was calculated based on the primary hypothesis and the ranked DOOR scores (accounting for ties). Assuming a 60% probability of a better DOOR in the dabigatran treatment group vs the oral FXaI treatment group, with at least 80% power, an alpha of 0.025, and a one-sided Wilcoxon rank sum test (dabigatran1. Secondary outcomes and the preplanned subgroup analysis are discussed in the Supplement. Data monitoring and safety SNAP is overseen by an independent DSMC. This phase 4 comparative effectiveness trial follows the guidelines of the International Conference on Harmonization Good Clinical Practice. All treatments being studied are known to be safe and effective for the underlying indication. The level of reporting of adverse events depends on the regulatory requirements in each region, but at a minimum, sites will report all serious adverse events related to the treatments being studied. We are capturing both bleeding and thrombotic events as key secondary outcomes. There is no planned interim analysis or early stopping criteria. The DSMC will make the final recommendation to stop the trial after the intended sample size has been achieved. Trial monitoring will take place and will follow a standard trial monitoring plan developed by the sponsor. Ethics This study has received approval from the McGill University Health Centre [study identifier: 2025-10900]. Written informed consent will be obtained from trial participants or an appropriate surrogate by trained research assistants. Any protocol amendments will be communicated in writing to trial investigators, the trial registry and to the DSMC; according to good clinical practice to trial participants; and according to local site regulations to the research ethics board. All patients will be followed according to usual care for SAB. Dissemination policy Trial participants will be offered a final copy of the study results sent to them electronically or via the mail. Study results will be disseminated at global conferences for infectious diseases (eg, IDSA, Infectious Diseases Society of America and ESCMID, European Society of Clinical Microbiology and Infectious Diseases). The study will be published open access in a high impact peer-reviewed journal and results will be disseminated via country and region-level infectious diseases associations. Principal investigators, site investigators and key study personnel will be offered the opportunity to meet International Committee of Medical Journal Editors (ICMJE) criteria for authorship and the manuscript will be written by the principal investigators without the aid of a medical writer. Access to the trial dataset and statistical analysis may be made available upon reasonable request and with an institutional data sharing agreement with the trial sponsor. Discussion In vitro data,10 11 animal models,12 observational studies14 15 and a small randomised controlled trial13 support the potential role of dabigatran alongside standard treatment for SAB. Despite this potential, there are countless examples where promising therapies based on good supporting observational evidence have not translated to clinical benefit. This nested RCT allows for the efficient generation of high-quality evidence to inform practice. A major advantage of this trial is that all patients have an established and approved indication for antithrombotic therapy that we are leveraging to answer relevant questions about SAB. In contrast to a study seeking to compare the therapeutic efficacy of newly initiated antithrombotic therapy to no antithrombotic therapy, this proposed trial minimises the bleeding and thrombotic risks assumed by subjects given that they would be administered a DOAC regardless of trial participation. From this data, we will be able to determine if dabigatran should be the preferred agent for adjunctive treatment among patients with SAB who require a DOAC, and a future phase 3 trial could then determine if the drug should be added de novo informed by the outcomes of this trial. Relatedly, pharmacologic P2Y12 inhibition with the molecule clopidogrel has also been shown to provide protection against lethal S. aureus challenge in murine models.23 Antiplatelet therapy in general may prevent platelet aggregation and binding to damaged endothelium (eg, heart valves). A multiyear observational study of SAB at Veterans Association hospitals in the USA reported that clopidogrel users were less likely to die in a propensity score-adjusted analysis (30-day mortality HR was 0.43 [95% CI 0.19 to 0.98]).24 The SNAP investigators will also be conducting a parallel study of the effect of aspirin vs clopidogrel on SAB outcomes among patients with an independent indication for treatment with aspirin. The protocol for this sister silo study will be published separately. Overall, the SNAP trials have heralded an unprecedented era in advancing the evidence-based management of SAB. At the time of writing, more than 4600 patients have been recruited to the platform contributing to the ongoing antibiotic domains.46 Ideally completed within 3–4 years, the Dabi-SNAP trial represents the first, but not the last, non-antibiotic therapy which will be tested on this revolutionary platform. Supplementary material