Bloodstream infections caused by Staphylococcus aureus carry a mortality burden that rivals many cancers, yet decades of clinical practice have rested on an unresolved debate: which antibiotic class should anchor treatment? A rigorous international trial now offers the clearest answer to date, with implications for how clinicians balance efficacy and organ safety in critically ill patients.
Conducted within a Bayesian adaptive platform trial across multiple countries, the study randomized adults with penicillin-resistant, methicillin-susceptible S. aureus (MSSA) bacteremia to receive either cefazolin or an antistaphylococcal penicillin — flucloxacillin or cloxacillin. The primary endpoint was all-cause mortality within 90 days, analyzed using a hierarchical Bayesian logistic-regression model. Enrollment ran from February 2022 to August 2024, closing once the prespecified noninferiority criterion — an adjusted odds ratio below 1.2, approximating less than 2.5 percentage points absolute mortality difference — was met. In the evaluable population, 90-day mortality reached 15.0% (97 of 645 patients) in the cefazolin arm, suggesting performance statistically on par with the comparator. A key secondary outcome tracked acute kidney injury within 14 days, a clinically meaningful safety signal given that antistaphylococcal penicillins carry well-documented nephrotoxicity risk.
This finding lands at a meaningful juncture in infectious disease medicine. For years, antistaphylococcal penicillins like flucloxacillin held a perceived pharmacodynamic advantage due to their narrower spectrum and historical precedent, while cefazolin — more widely available globally, easier to administer, and potentially gentler on the kidneys — was viewed with hesitation in severe bacteremia. Earlier observational data hinted at cefazolin's competitiveness, but were limited by confounding and selection bias. A randomized noninferiority result from a Bayesian platform trial — a design well-suited to adaptive evidence accumulation in complex infections — substantially elevates the evidence grade. That said, the open-label design introduces potential for treatment-level biases, and the platform's ongoing nature means certain subgroup effects may yet be characterized. Whether superiority in renal outcomes translates to meaningful clinical advantage at the population level will require full data disclosure. This is confirmatory and potentially practice-shifting, particularly for health systems where antistaphylococcal penicillins are scarce or nephrotoxicity risk is elevated.