Background Methicillin-susceptible Staphylococcus aureus (MSSA) bacteraemia is a common and serious infection, associated with significant morbidity and high mortality rates. Antistaphylococcal penicillins (ASPs) have traditionally been recommended as the first-line treatment. However, this established position has recently been challenged by meta-analyses suggesting that cefazolin may provide comparable efficacy, along with a more favourable safety profile. Further clinical and real-world studies are warranted to substantiate these findings.

Objectives The aim of this study was to assess the effectiveness and safety of cefazolin compared with ASPs in the management of MSSA bacteraemia.

Staphylococcus aureus (SA) is a frequent cause of bacteraemia and is associated with high mortality rates. Several studies have reported an increasing incidence over recent decades, accompanied by a decline in the prevalence of methicillin-resistant strains. In a recent meta-analysis by Bai et al., mortality related to S. aureus bacteraemia (SAB) remained substantial: 18.1% at one month, 27.0% at three months, and 30.2% at one year.

In the absence of specific international guidelines supported by randomised controlled trials, the management of methicillin-susceptible S. aureus (MSSA) bloodstream infections largely relies on expert consensus and clinical experience. For several decades, antistaphylococcal penicillins (ASPs) have been regarded as the gold standard for treating MSSA bacteraemia. This preference was primarily based on concerns regarding the so-called "inoculum effect," whereby the efficacy of cefazolin might be reduced at high bacterial loads .

However, recurring global shortages of ASPs have led clinicians to consider cefazolin as a viable first-line alternative. In this context, several meta-analyses, albeit based on observational data, have suggested that cefazolin provides similar efficacy with a more favourable safety profile than ASPs . Nonetheless, these results remain subject to confounding and bias and must be confirmed through randomised controlled trials, two of which are currently ongoing.

Large, real-world evaluations are therefore needed to assess the comparative effectiveness and safety of cefazolin and ASPs in routine clinical settings. The TriNetX platform offers access to real-time, real-world global hospital data, thereby enabling large-scale comparative analyses in an international context. Against this background, the present study aimed to evaluate the effectiveness and tolerability of cefazolin versus ASPs in the treatment of MSSA bacteraemia using data from the TriNetX database.

Inclusion Criteria:

• Monobacterial methicillin susceptible Staphylococcus aureus (MSSa) bacteremia
• In the cefazolin group : Cefazolin injectable prescription in the week before or the week after MSSa bacteremia
• In the ASPs group :ASPs injectable prescription (flucloxacillin, floxacillin, cloxacillin, oxacillin, nafcillin and dicloxacillin) in the week before or the week after MSSa bacteremia

Exclusion Criteria:

• In the cefazolin group : ASP injectable prescription in the year before or the year after MSSa bacteremia
• In the ASPs group : cefazolin injectable prescription in the year before or the year after MSSa bacteremia

Variables : age, sex, medical history, type of infection and antibiotic used.

Variables will be described as number and percentage for categorical variables, and mean and standard-deviation for continuous ones. A propensity score matching method will be conducted including (i) variables significantly different between the two cohorts in the bivariate descriptive analyses between the patient birth and the index event date (standard difference < 0.10 ; appendix 1), and (ii) clinically relevant variables (known confusion factors linked to mortality and safety outcomes ; appendix 2) at the index event date. Variables related to patients will be described before and after propensity score matching.

Data are available upon reasonable request. Researchers can request access to data from the TriNetX research network through the TriNetX platform (https://live.trinetx.com). However, this may involve associated costs, require a data-sharing agreement and no patient-identifiable information can be accessible.