The optimal therapy for methicillin-susceptible Staphylococcus aureus bacteremia has been a subject of healthy debate in recent years. Anti-staphylococcal penicillins such as nafcillin and oxacillin have traditionally been considered as treatments of choice, but cefazolin has demonstrated similar efficacy in multiple observational studies with the advantages of less frequent dosing and lower rates of adverse drug events, including hypersensitivity and renal impairment. One caveat, however, is that there have been reports of cefazolin failure in the treatment of high-inoculum infections caused by isolates that overproduce the blaZ beta-lactamase (mostly type A blaZ).
A study recently published in The Lancet addresses this controversy. The authors randomized 315 patients at 21 hospitals in France with MSSA bacteremia to receive their first 7 days of treatment with either cloxacillin or cefazolin, with subsequent choice of therapy (for at least 14 days total) at the discretion of the participating investigator. The primary endpoint — a composite of sterile blood cultures at day 3 (or day 5 for endocarditis) without bacteremia relapse, survival and clinical success at day 90 — was met for 74% of the patients receiving cloxacillin and 75% in those receiving cefazolin, with significantly less serious adverse events in the cefazolin group (15% vs. 27%, P = 0.010), in particular, less kidney injury (1% vs. 12%).
The authors also determined blaZ status of all isolates: 70% carried blaZ, similarly distributed between types A, B and C; 33% of all blaZ were type A. Noninferiority of cefazolin over cloxacillin was met for all outcomes among all those without type A blaZ isolates but not met for all outcomes among those with type A blaZ (though sample size was limited).
All in all, cefazolin proved to be noninferior to cloxacillin for up-front management of MSSA bacteremia. However, the signal toward lack of noninferiority among patients with isolates carrying type A blaZ likely deserves further attention and study, and the notion of adding blaZ type to rapid molecular diagnostic panels for blood culture identification should be explored.
(Burdet et al. Lancet. 2025;406(10517):2349-2359).)