Cefazolin matches older penicillins for staph bloodstream infection, with less kidney harm

Staphylococcus aureus in the bloodstream is one of the most dangerous common infections in hospital medicine, and it kills a substantial share of the people who get it [s1]. For the methicillin-susceptible form — the strains that ordinary anti-staph antibiotics can still treat — clinicians have argued for years over which first-line drug is better: cefazolin, a cephalosporin, or an antistaphylococcal penicillin such as flucloxacillin or cloxacillin. A large randomised comparison has now weighed them directly, and the answer is pragmatic rather than dramatic. Cefazolin was non-inferior for survival, and it spared the kidneys [s1].

The comparison came from SNAP, an international Bayesian adaptive platform trial — a design that runs several treatment questions at once and lets evidence accumulate until a prespecified threshold is crossed [s1]. Within it, investigators conducted an open-label, randomised comparison of cefazolin against an antistaphylococcal penicillin in adults with penicillin-resistant, methicillin-susceptible S. aureus bacteraemia [s1]. The primary outcome was death from any cause within 90 days of enrolment, assessed with a hierarchical Bayesian logistic-regression model [s1].

How the trial framed a tie

Non-inferiority trials ask a specific question: not "is the new drug better?" but "is it at least not meaningfully worse?" SNAP set that bar in advance. Cefazolin would be judged non-inferior if the adjusted odds ratio for death was below 1.2 — a margin that approximates an absolute difference in mortality of less than 2.5 percentage points, assuming mortality in the penicillin group of 15% [s1]. The trial also tested for outright superiority, defined as an adjusted odds ratio below 1.0 [s1]. Framing both thresholds ahead of time is what keeps a "tie" from being reverse-engineered after the numbers land. The trial is registered as NCT05137119 [s2].

What it showed

This domain of the ongoing trial ran between 17 February 2022 and 7 August 2024, by which point the non-inferiority criterion had been met [s1]. Mortality at 90 days among adults who could be evaluated was 15.0% — 97 deaths among 645 patients — in the cefazolin group, against 17.0% — 109 deaths among 642 patients — in the antistaphylococcal-penicillin group [s1]. The adjusted odds ratio was 0.81 (95% credible interval 0.59 to 1.12) [s1]. In the language of the design, the posterior probability that cefazolin was non-inferior was 99.2%, and the probability that it was actually superior was 89.8% [s1] — strong support for non-inferiority, and a lean toward benefit that fell short of the trial's stricter superiority bar.

The clearer separation came on safety. Acute kidney injury within 14 days occurred in 92 of 660 patients (13.9%) in the cefazolin group, compared with 127 of 648 (19.6%) in the penicillin group — an adjusted odds ratio of 0.67 (95% credible interval 0.50 to 0.89), with a 99.7% probability of superiority [s1]. Kidney injury is a recognised hazard of the antistaphylococcal penicillins, and here cefazolin measurably lowered it. For a patient already fighting a bloodstream infection, avoiding a second organ insult is not a minor consideration — acute kidney injury lengthens hospital stays, complicates further drug dosing, and is itself linked to worse survival.

The limits

The comparison was open-label — patients and treating teams knew which drug was given — which can bias softer judgements, though death is a hard, unambiguous endpoint less vulnerable to that. The population was specifically penicillin-resistant, methicillin-susceptible S. aureus, so the result speaks to that group rather than to every staph bacteraemia. Cefazolin also carries a known theoretical vulnerability, the "inoculum effect," in which very high bacterial loads can blunt the drug; this trial's overall result did not show that translating into worse survival, but it does not settle every subgroup. And because SNAP is a platform trial that stopped this domain once its threshold was crossed, the estimate reflects an efficient stopping point rather than a fixed enrolment target.

What to watch

SNAP was funded by the National Health and Medical Research Council and others, and sponsored by the University of Melbourne [s1][s2] — an academically led, publicly funded trial rather than a manufacturer's study, which matters when the finding may shift routine prescribing away from long-standing habit. The investigators' conclusion is measured: in methicillin-susceptible S. aureus bacteraemia, cefazolin was non-inferior to flucloxacillin or cloxacillin for 90-day mortality and was associated with less acute kidney injury [s1]. That combination — equivalent survival, fewer kidney complications — is the kind of evidence that tends to move guidelines. What remains to be watched is whether the platform's further domains, and real-world use across different strain populations, hold the same line.

Sources

• [s1] Cefazolin for Methicillin-Susceptible Staphylococcus aureus Bacteremia. New England Journal of Medicine. 17 June 2026. doi:10.1056/NEJMoa2506905. PMID 42308484.
• [s2] ClinicalTrials.gov. Staphylococcus aureus Network Adaptive Platform Trial (SNAP), NCT05137119. U.S. National Library of Medicine.