Abstract
Background
Whether cefazolin or an antistaphylococcal penicillin should be preferred for treating methicillin-susceptible Staphylococcus aureus (MSSA) bacteremia is unclear.
Methods
In an ongoing, international, Bayesian adaptive platform trial, we conducted an open-label, randomized comparison of cefazolin versus an antistaphylococcal penicillin (flucloxacillin or cloxacillin) in adult patients with penicillin-resistant, methicillin-susceptible S. aureus bacteremia. The primary outcome was 90-day all-cause mortality, evaluated using a hierarchical Bayesian logistic regression model. We report the posterior probability of non-inferiority (pre-specified as an adjusted odds ratio [aOR] < 1.2, approximating < 2.5% absolute difference if mortality in the (flu)cloxacillin arm was 15%) and of superiority (corresponding to an aOR <1.0). Secondary safety outcomes included the development of acute kidney injury within 14 days.
Results
This trial domain was conducted between February 17, 2022, and August 7, 2024, at which time non-inferiority was met. Death at 90 days amongst evaluable adults was 15.0% (97/645) for cefazolin and 17.0% (109/642) for (flu)cloxacillin (aOR 0.81; 95% Credible Interval [CrI], 0.59–1.12; probability of non-inferiority 99.2% and superiority 89.8%). Acute kidney injury occurred in 92/660 (13.9%) patients in the cefazolin arm versus 127/648 (19.6%) in the (flu)cloxacillin arm (aOR 0.67, 95%CrI 0.50–0.89; probability of superiority 99.7%).
Conclusion
In patients with MSSA bacteremia, cefazolin had a high probability of non-inferiority to (flu)cloxacillin for 90-day mortality, with less acute kidney injury. (Funded by National Health and Medical Research Council and others; Clinicaltrials.gov NCT05137119).
S. aureus bacteremia is a leading cause of bacterial-related mortality worldwide1 with more than 1 in 4 afflicted patients dying within 90 days.2 While there are accepted best practices in the management of S. aureus bacteremia,3 the preferred antibiotic for treating methicillin-susceptible S. aureus (MSSA) is unclear.4 Expert opinion from endocarditis guidelines5,6 favors the antistaphylococcal penicillins (e.g., nafcillin, oxacillin, cloxacillin, or flucloxacillin) over cefazolin5 due to theoretical concerns about the “cefazolin inoculum effect”, an in vitro destruction of cefazolin by beta-lactamases which may be relevant in high burden infections.7,8 However, meta-analyses of observational studies suggest that cefazolin could be superior in terms of 30-day mortality (odds ratio 0.73; 95%CI 0.62-0.85)9 with a more favorable adverse event profile when compared to antistaphylococcal penicillins. Given observational analyses may be subject to multiple biases,10 a randomized clinical trial is the most reliable way to inform patient management.
We launched the S. aureus Network Adaptive Platform (SNAP) trial in 2022 to answer multiple questions related to the treatment of patients with S. aureus bacteremia.11 Within this Bayesian adaptive platform trial, we conducted a pragmatic, open-label, randomized comparison of cefazolin to an antistaphylococcal penicillin (cloxacillin or flucloxacillin depending on the country [hereafter (flu)cloxacillin]) for the treatment of MSSA bacteremia.
Methods
Study Design and Setting
SNAP contains a series of investigator-initiated, international, multicenter studies.11 The platform features a master protocol and integrated statistical framework12 with domain specific appendices containing nested questions. Healthcare consumers were involved in trial design through focus groups and representation on steering committees. Each domain11 refers to a management question with at least two interventions being compared (e.g., backbone antibiotics, adjunctive clindamycin therapy13, or early oral switch14; see Supplementary Appendix). The backbone domain is further divided into silos based on S. aureus antibiotic susceptibility: penicillin-susceptible (PSSA); methicillin-susceptible, penicillin-resistant (MSSA); and methicillin-resistant (MRSA).
Herein we report the open-label, parallel group, non-inferiority, randomized comparison of patients in the MSSA silo of the antibiotic backbone domain of SNAP. Participants in this domain and silo were recruited between February 17, 2022, and June 21, 2024, at 91 sites in 8 countries (Australia, Canada, Israel, New Zealand, the Netherlands, Singapore, South Africa, and the United Kingdom; see Supplementary Appendix) with the last participant reaching 90-days of follow-up on September 21, 2024. Ethics and regulatory approval were obtained at each participating center. Written or oral informed consent, in accordance with regional regulations, was obtained from all patients and/or their surrogates. The study was designed by the investigators via international working groups. Data was analyzed by a dedicated analytic team of investigators. Tong, Norton, Mahar, McGlothlin, and Dymock vouch for the data and analysis. The study was supported by funding organizations and academic sponsors in multiple countries (Supplementary Appendix) who had no role in the design, analysis, or reporting of the results. The master protocol,11 statistical methods,12 domain specific appendix, and statistical analysis plan are available in the Supplementary Appendix and the protocol at nejm.org. SNAP is registered on clinicaltrials.gov (NCT05137119).
Participants
We enrolled hospitalized patients with S. aureus bacteremia to the platform within 72 hours of their index blood culture collection. This report describes adults (≥18 years). The pedia tric s tudy is ongoing. Platform and domain eligibility criteria are detailed in the Supplementary Appendix. Notable platform exclusions included enrollment beyond 72 hours from collection of the first positive blood culture, polymicrobial bacteremia, or patients not for active treatment. Key domain specific exclusions included an active history of penicillin or cefazolin allergy; current receipt of maintenance dialysis; and ongoing treatment with other systemic antibacterial agents with S. aureus activity. Additionally, because the cefazolin inoculum effect is dependent on the production of penicillinase, we only included penicillin-resistant MSSA (see Laboratory Methods) to avoid bias towards non-inferiority of cefazolin.
Randomization and Allocation
All participants were randomized 1:1 within the MSSA silo at platform entry by simple randomization without stratification using a web-based interactive randomization system (Spiral Software, New Zealand). As soon as participants were confirmed to be eligible for the MSSA silo of the backbone domain, their randomized allocation to cefazolin or (flu)cloxacillin was revealed.
Interventions
Recommended dosing was: 2g intravenously (IV) q8h for cefazolin (q6h for critical illness or endocarditis); 2g IV q6h for flucloxacillin (q4h for critical illness or endocarditis), and 2g IV q4h for cloxacillin, with dose adjustments for renal function (Supplementary Appendix). Other than potential treatment assignment in the adjunctive antibiotic domain,13 post-enrollment administration of additional antibiotics was discouraged.
Infectious diseases consultation was advised. The protocol recommended allocated antibiotics be continued for the entire intravenous treatment duration, with a minimum antibiotic duration of 14 days for uncomplicated bacteremia and 28–42 days for complicated bacteremia.15 Subsequent randomization to the oral switch domain14 was permitted at 7 and 14 days, respectively for eligible participants. Apart from the antistaphylococcal antibiotics used, clinical management was left to the treating clinician’s discretion.
Outcomes
The primary outcome for the platform and domain was all-cause mortality at 90 days after platform entry. Secondary platform outcomes included: all-cause mortality at 14, 28, and 42 days; microbiologic treatment failure and diagnosis of new foci of infection between days 15 and 90; and Clostridioides difficile infection within 90 days. Secondary domain specific outcomes (further defined in Supplementary Appendix) included acute kidney injury (AKI), defined as an absolute creatinine increase ≥26.5 μmol/L (0.3mg/dL) within 5 days or a relative increase ≥50% from baseline within the first 14 days, initiation of renal replacement therapy, and hepatotoxicity.
Adverse Events
In addition to the pre-specified safety outcomes, site investigators were asked to conform to regional requirements and, at a minimum, record serious adverse reactions defined as all serious adverse events that were thought to be possibly, probably, or definitely related to study drugs or study procedures.
Laboratory Methods
Methicillin and penicillin susceptibility were determined locally using automated testing methods and/or phenotypic disc diffusion16 and/or molecular testing for the mecA and blaZ genes (Supplementary Appendix).
Statistical Methods
Study Populations
Populations were defined based on the principle of intention-to-treat (ITT) for each estimand corresponding to the revealed randomized treatment assignment, regardless of receipt. The primary analysis was also performed on a protocol-adherent population (Supplementary Appendix).
Analyses
The scheduled and final analyses were performed by an independent unblinded analytic team which was separate from the trial team. For this analysis, we used a minimally-informative prior for all model parameters. The analyses are described in the statistical appendix included with the Supplementary Appendix.
The adjusted odds ratio (aOR) for the primary outcome was calculated using a hierarchical Bayesian logistic regression model. This model incorporated data from all platform participants to adjust for age, country, temporal epoch (contiguous 26-week periods from the start of the domain), and eligibility and treatment assignment in other platform domains. Details of the model fitting and diagnostics are provided in the Supplementary Appendix. The treatment effect pertaining to cefazolin versus (flu)cloxacillin was calculated exclusively by using patients in this silo’s randomized comparison (Figure 1) with pre-specified information sharing from the 87 pediatric participants enrolled in this silo (study ongoing). The probability of non-inferiority (pre-specified as an aOR<1.2; approximating an absolute margin of 2.5% absolute difference if mortality in the [flu]cloxacillin arm was 15%) and superiority (aOR<1.0) were estimated. The primary analysis for the primary outcome excluded participants with missing 90-day mortality data (complete case analysis). Post hoc analyses included analyses where all participants whose day 90 mortality data was missing were assumed to have survived or to have died.
Figure 1. Enrolment, randomization, and follow-up.
Notes: MSSA denotes methicillin-susceptible Staphyloccocus aureus.
Binary secondary outcomes were analyzed similarly. We planned a time-to-event analysis using a Bayesian Weibull proportional hazards model; however, this model’s assumptions were not satisfied and so we present unadjusted Kaplan-Meier curves. Pre-specified subgroup analyses for the primary outcome were the presence or absence of endocarditis and the presence or absence of suspected/proven central nervous system (CNS) infection. During peer review, a post hoc subgroup analysis was conducted for a shorter and longer duration of pre-treatment (≤48 hours, 48-72 hours).
As analyses were conducted in a Bayesian framework, we report the means of the posterior distributions of treatment effects with 95% credible intervals and where appropriate associated posterior probabilities. No P-values were calculated, and no adjustment for multiple analyses was pre-specified. Therefore, for the secondary estimands, excluding safety outcomes, the 95% credible intervals should not be used to draw definitive conclusions.
Interim Analyses and Decision Rules
As a perpetual adaptive Bayesian platform, simulations were conducted to optimize trial operating characteristics such as number and timing of interim analyses, decision rule thresholds, and with a maximum ceiling sample size of 6,000 patients.12 We planned and conducted scheduled analyses after every 500 platform patients reached day 90 with pre-specified stopping rules at each analysis for each domain.12 For the MSSA silo of the backbone domain, if the posterior probability of non-inferiority for cefazolin was greater than 99% in adults, the independent data and safety monitoring committee (DSMC) could recommend stopping recruitment. If the posterior probability of superiority for adults was less than 1%, the DSMC could recommend stopping for futility. If no threshold was met, recruitment was to continue unless the DSMC had substantial safety concerns which were not prespecified. For the MSSA non-inferiority comparison, simulations indicated an approximate frequentist type I error of 2%.
On June 21, 2024, at the fourth planned interim analysis, the DSMC requested recruitment be paused in the MSSA silo because of a safety signal involving AKI. Subsequently, on August 5, 2024, they recommended closing the MSSA silo because the statistical trigger for non-inferiority had been met, and the safety signal persisted. On August 7, 2024, without additional knowledge on study outcomes, the trial steering committee accepted the recommendation.
Study Oversight
International coordination was provided by the University of Melbourne with the direction of a global trial steering committee.11 Within each participating region, study coordination and monitoring was provided by a sponsor who was responsible under each region’s laws and governance.
Results
Study Population
Of 8546 patients screened as of June 21, 2024, 2602 adults (30.4%) were recruited to the platform and underwent one or more randomizations (Fig. S1). Of those, 1757 had MSSA (67.5%) and 1341/1757 (76.3%) participants were randomized in the MSSA silo (671 to cefazolin, 670 to (flu)cloxacillin; Figure 1). Fifty-four of these patients (4.0%) were lost to follow-up, leaving 1287 patients in the primary outcome analysis. Recruitment to SNAP is ongoing and to maintain trial integrity, the results for the pediatric patients, MRSA silo, and other domains remain blinded.
Baseline characteristics are presented by treatment assignment in Table 1. The median age was 66 years (interquartile range [IQR] 53–76) and 421 (31.4%) were female, which are representative of a typical cohort with S. aureus bacteremia (Tables S1-S3). The most common site of infection was osteoarticular (32.1%) (Table 2). Antibiotics received prior to randomization allocation are presented in Table S4. Four hundred and eighty-eight patients (36.4%) were receiving cefazolin at enrollment and 665 (49.6%) (flu)cloxacillin, with similar numbers across the assigned treatment groups. More than 99% of patients in this study received infectious diseases consultations. Both treatment groups received a similar duration of intravenous and oral antibiotics (Table S5).
Table 1. Baseline Characteristics of Patients.
| No./Total No.(%) | ||
|---|---|---|
| Variable | Cefazolin (N=671) | (Flu)cloxacillin (N=670) |
| Demographics | ||
| Median age (IQR) | 66 (53−76) | 66 (52−77) |
| Female sex | 215 (32.0) | 206 (30.7) |
| Median weight in kg (IQR) | 82.0 (69.0−96.8) | 80.0 (68.0−95.0) |
| Comorbidities | ||
| Diabetes mellitus | 231 (34.4) | 217 (32.4) |
| Chronic kidney disease1,2 | 96 (13.9) | 93 (14.3) |
| Cirrhosis | 34 (5.1) | 30 (4.5) |
| Cancer requiring any surgical or medical therapy within 12 months3 | 95 (14.2) | 76 (11.3) |
| Dementia | 23 (3.4) | 32 (4.8) |
| Iatrogenic immunosuppression4 | 105 (15.6) | 84 (12.5) |
| Implanted intravascular prosthesis or endovascular device5 | 64 (9.5) | 58 (8.7) |
| Previous infective endocarditis | 11 (1.6) | 10 (1.5) |
| Underlying cardiac abnormality predisposing to endocarditis6 | 50 (7.5) | 42 (6.3) |
| Injection drug use within 6 months | 31 (4.6) | 52 (7.7) |
| Limitations to care7 | 80 (11.9) | 84 (12.5) |
| Prognostic Factors | ||
| Median hours to index blood culture positivity (IQR) | 15.7 (12.0−21.0) | 15.3 (12.0−20.7) |
| Median hours from index blood culture to platform entry (IQR) | 52.3 (43.5−65.5) | 50.9 (42.3−64.7) |
| Median hours from index blood culture to silo entry (IQR) | 62.2 (50.2−67.9) | 61.1 (50.0−67.5) |
| Intensive care unit at time of recruitment | 74 (11.0) | 72 (10.7) |
| Median Pitt bacteremia score (IQR) | 0 (0−1) | 0 (0−1) |
| Median C-reactive protein (mg/L) (IQR) | 192.0 (113.5−266.5) | 203.1 (122−283.8) |
Defined as estimated glomerular filtration rate below 50mL/min/1.72m^2;
Four patients (0.3%) were receiving baseline hemodialysis or peritoneal dialysis (protocol deviations). They were retained in the intention-to-treat population
Excluding non-melanoma skin cancers
Within past 3 months has received any of the following: prednisone >0.5mg/kg/day (or equivalent) for more than 14 days, cyclosporine, tacrolimus, or sirolimus, azathioprine, leflunomide, or mycophenolate, cyclosporine or methotrexate, monoclonal antibodies such as infliximab, cancer chemotherapy, bone marrow transplantation, other immunosuppressive agents equal to above, AIDS
Not including central lines whether temporary or tunneled
Excluding prosthesis above
Active orders or advance care directives limiting resuscitation and intensive care admission.
Table 2. Focus of Infection.
| No./Total No.(%) | ||
|---|---|---|
| Focus of Infection1 | Cefazolin (N=671) | (Flu)cloxacillin (N=670) |
| Osteoarticular (including epidural abscess) | 220 (32.8%) | 210 (31.3%) |
| Septic arthritis and extra-axial osteomyelitis | 167 (24.9%) | 151 (22.5%) |
| Vertebral osteomyelitis and/or septic discitis | 61 (9.1%) | 59 (8.8%) |
| Epidural abscess | 45 (6.7%) | 49 (7.3%) |
| Orthopedic hardware-associated | 34 (5.1%) | 31 (4.6%) |
| Skin and soft tissue | 177 (26.4%) | 204 (30.4%) |
| Isolated skin and soft tissue | 127 (18.9%) | 136 (20.3%) |
| Intravascular catheter associated | 103 (15.4%) | 102 (15.2%) |
| Isolated intravascular catheter infection | 91 (13.6%) | 91 (13.6%) |
| Primary bacteremia (no other focus identified) | 81 (12.1%) | 80 (11.9%) |
| Endocarditis | 55 (8.2%) | 57 (8.5%) |
| Left-sided native valve | 35 (5.2%) | 33 (4.9%) |
| Right-sided native valve | 13 (1.9%) | 19 (2.8%) |
| Prosthetic valve | 13 (1.9%) | 9 (1.3%) |
| Pleuropulmonary | 40 (6.0%) | 42 (6.3%) |
| Other deep focus | 27 (4.0%) | 37 (5.5%) |
| Genitourinary | 24 (3.6%) | 22 (3.3%) |
| Native vascular or vascular graft | 15 (2.2%) | 16 (2.4%) |
| Unspecified | 14 (2.1%) | 13 (1.9%) |
| Cardiac or other implantable non-orthopedic device | 12 (1.8%) | 16 (2.4%) |
| Central Nervous System | 5 (0.7%) | 2 (0.3%) |
Patients may have had more than one focus of infection except where isolated line and skin and soft tissue infection is specified
Primary Outcome
Ninety-seven of 645 evaluable patients (15.0%) in the cefazolin group and 109 of 642 evaluable patients (17.0%) in the (flu)cloxacillin group met the primary outcome of mortality at 90 days (aOR 0.81; 95% CrI 0.59–1.12) corresponding to a posterior probability of non-inferiority of 99.2% and superiority of 89.8% (Table 3, Fig. S2). In the 3.8% of patients with missing primary outcome data, the baseline characteristics were similar to those with complete data, apart from a younger age and history of injecting drug use in both treatment arms (Table S6). The complementary analyses accounting for missing outcomes were congruent with the primary analysis having posterior probabilities of non-inferiority exceeding 99%. A further post hoc sensitivity analysis using a broader prior demonstrated similar results to the primary analysis (Table S7).
Table 3. Primary and Secondary Outcomes.
| No./Total No.(%) | |||||
|---|---|---|---|---|---|
| Outcome or Analysis | Cefazolin (N=671) |
(Flu)cloxacillin (N=670) |
Median Adjusted Odds Ratio (95%Credible Interval) |
Posterior Probability of Cefazolin Non-Inferiority |
Posterior Probability of Cefazolin Superiority |
| Primary Outcome - 90 Day Mortality | |||||
| Primary Analysis Population | 97/645 (15.0%) | 109/642 (17.0%) | 0.81 (0.59–1.12) | 99.2% | 89.8% |
| Assuming all missing were dead | 123/671 (18.3%) | 137/670 (20.4%) | 0.83 (0.63–1.10) | 99.6% | 90.7% |
| Assuming all missing were alive | 94/671 (14.5%) | 109/670 (16.3%) | 0.81 (0.59-1.11) | 99.3% | 90.5% |
| Protocol-adherent Population | 81/589 (13.8%) | 73/520 (14.0%) | 0.88 (0.61–1.26) | 95.4% | 76.0% |
| Secondary Outcomes | |||||
| Acute Kidney Injury within 14 days1 | 92/660 (13.9%) | 127/648 (19.6%) | 0.67 (0.50–0.89) | >99.9% | 99.7% |
| Hepatotoxicity within 14 days2 | 75/574 (13.1%) | 78/562 (13.9%) | 0.96 (0.68–1.35) | 90.0% | 59.7% |
| Any serious adverse reaction related to study drug | 12/671 (1.8%) | 33/670 (4.9%) | 0.41 (0.21–0.77) | 99.9% | 99.8% |
| Day 14 Mortality | 25/667 (3.7%) | 37/665 (5.6%) | 0.67 (0.39–1.11) | NR | NR |
| Day 28 Mortality | 47/665 (7.1%) | 70/665 (10.5%) | 0.61 (0.41–0.91) | NR | NR |
| Day 42 Mortality | 63/663 (9.5%) | 86/662 (13.0%) | 0.66 (0.46–0.94) | NR | NR |
| Change of antibiotic due to perceived inefficacy | 14/671 (2.1%) | 16/670 (2.4%) | 1.05 (0.51–2.10) | NR | NR |
| Microbiologic Treatment Failure After Day 14 | 18/587 (3.1%) | 12/566 (2.1%) | 1.41 (0.71–2.78) | NR | NR |
| Diagnosis of New Foci of Infection after Day 14 | 28/623 (4.5%) | 30/609 (4.9%) | 0.87 (0.53–1.45) | NR | NR |
| Change of antibiotic due to adverse event | 11/671 (1.6%) | 61/670 (9.1%) | 0.21 (0.11–0.38) | NR | NR |
| New Renal Replacement Therapy within 90 days | 17/668 (2.5%) | 27/657 (4.1%) | 0.61 (0.33–1.11) | NR | NR |
| Ongoing Renal Replacement Therapy at 90 days | 4/662 (0.6%) | 3/643 (0.5%) | 1.00 (0.31–3.16) | NR | NR |
| C. difficile infection3 | 14/664 (2.1%) | 10/661 (1.5%) | 1.28 (0.60–2.66) | NR | NR |
| Intravenous catheter complication requiring removal4 | 22/668 (3.3%) | 35/665 (5.0%) | 0.68 (0.39–1.16) | NR | NR |
Defined as a creatinine increase ≥26.5 μmol/L (0.3mg/dL) or ≥50% from baseline
Defined as alanine transaminase or gamma-glutamyl transferase increasing above 2.5 times the upper limit of normal
Defined as having symptoms, a positive test, and receiving therapy
While on assigned therapy.
Posterior probabilities are not reported for secondary outcomes, with the exception of safety outcomes, and 95% credible intervals do not include adjustment for multiplicity.
Secondary Outcomes
Secondary outcomes are presented in Table 3. At every time point prior to day 90, cefazolin had a probability of non-inferiority for mortality exceeding 98.8% and a probability of superiority exceeding 94.1%. Unadjusted Kaplan-Meier plots for mortality are provided in Figure S3.
Patients receiving cefazolin experienced less AKI than those receiving (flu)cloxacillin (92/660 [13.9%] versus 127/648 [19.6%]; aOR 0.67; 95%CrI 0.50–0.89; probability of superiority 99.7%); including a 94.6% posterior probability of a reduced risk of initiating renal replacement therapy within 90 days (Table 3, Fig. S4). The staging of AKI is presented in the Table S9.
Protocol-adherent and Prespecified Subgroup Analyses
The protocol-adherent analysis demonstrated a 95.4% probability of cefazolin being non-inferior to (flu)cloxacillin (Table 3). Mortality for patients with endocarditis was 8/51 (15.7%) for cefazolin and 13/53 (24.5%) for (flu)cloxacillin (aOR 0.54; 95%CrI 0.20–1.43; probability of non-inferiority 94.4% and superiority 88.9%). There were only 7 patients with a central nervous system infection, precluding meaningful comparison.
Reported Adverse Events
Serious adverse reactions related to study drugs are detailed in Table 3 and Table S10. Overall, cefazolin had substantially lower odds of a serious adverse reaction being reported (aOR 0.41; 95%CrI 0.21–0.77) and of having to be discontinued due to adverse events (aOR 0.21; 95%CrI 0.11–0.38).
Discussion
In this pragmatic randomized clinical trial in adults with MSSA bacteremia, cefazolin was non-inferior to (flu)cloxacillin for the primary outcome of all-cause mortality at 90 days. Cefazolin also had lower risk of acute kidney injury; initiation of renal replacement therapy; and of treatment discontinuation due to adverse events. These data also suggest that cefazolin may reduce mortality, with an estimated probability of superiority of 94–99% at days 14, 28, and 42 which remained approximately 90% by day 90.
The adjusted odds ratio for 90-day mortality of 0.81 (CrI 0.59–1.12) for cefazolin versus (flu)cloxacillin is similar to pooled estimates from observational studies of an overall odds ratio of 0.73 compared to the antistaphylococcal penicillins and 0.31 to 0.92 for individual drugs within the class.9 While patients with S. aureus bacteremia are at direct risk of acute kidney injury by virtue of acute illness,17 the potential nephrotoxicity of the antistaphylococcal penicillins has been recognized since the 1960s.18 While we used flucloxacillin and cloxacillin in this trial, other antistaphylococcal penicillins have similar antibacterial activity and chemical structures,19 and based on observational data are likely to confer a similar risk of nephrotoxicity or treatment discontinuation for adverse events when compared to cefazolin.9 Therefore, the mortality and nephrotoxicity findings reported here may apply to other antistaphylococcal penicillins other than (flu)cloxacillin.
Cefazolin has some potential disadvantages compared to (flu)cloxacillin, including higher drug acquisition costs in some markets, and a potentially broader antimicrobial spectrum. However, these concerns are balanced by its longer half-life, reduced toxicity, and possible superiority to (flu)cloxacillin.
Limitations
For reasons of practicality and feasibility, this was an open-label trial; however, the primary outcome, all-cause mortality at 90 days, was objective and the decision criteria to stop the trial was based on pre-specified statistical rules. We had limited control over physician decisions post-randomization. In the presence of persistent bacteremia or other clinical factors, physicians could have done more diagnostic testing, changed or added antibiotics, or performed more aggressive source control. We did not collect and therefore cannot present these data. We donot believe post-randomization care would have been different by knowledge of treatment assignment, but this is a limitation of the open-label design and also that variability of practices across sites may have introduced heterogeneity. Fortunately, a change of assigned therapy due to clinician perceived treatment failure was less than 2.5% and the results of the protocol-adherent population analysis were consistent with the primary analysis. Second, we have not yet tested isolates for the presence of the cefazolin inoculum effect and will report this microbiological subgroup when testing of all isolates has been performed at central reference laboratories. In the interim, concerns about the clinical relevance of the cefazolin inoculum effect7 should be substantially allayed by the narrow non-inferiority margin and similar estimated treatment effect in patients with infective endocarditis compared to the overall population. Third, we presented a complete case analysis for the primary outcome (3.8% missing), with complementary analyses where all missing patients were considered either alive or dead yielding near identical results.
Conclusion
This pragmatic international, open-label platform trial demonstrated that cefazolin was non-inferior to (flu)cloxacillin for 90-day all-cause mortality in the treatment of adults with MSSA bacteremia. Cefazolin was associated with fewer serious adverse reactions, in particular less acute kidney injury.
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.
Supplementary Material
Acknowledgements
The authors wish to acknowledge the patients who volunteered to participate in the SNAP platform and their treating clinicians, without whom this work would not have been possible.
Funding
Funders had no role in the conduct or analysis of this RCT. The SNAP trial was funded by the National Health and Medical Research Council (1184238, 2014900, 2032628) and Medical Research Future Fund (2017301) in Australia; by the Canadian Institutes of Health Research (433304, 466322, 474605) and the Accelerating Clinical Trials Consortium (467903) in Canada; by UMC Utrecht and ZonMW Good Use of Medicines Program (10140022110014) in the Netherlands; by the Health Research Council of New Zealand (HRC Project Grant 20/344) and the Starship Foundation (ASF2144_WEBB) in New Zealand; by the National Medical Research Council (CTGIIT21nov-0002) in Singapore; the National Institute of Health and Care Research (NIHR133719, NIHR304258) and the Medical Research Council (MC_UU_00004/05) in the United Kingdom and by the National Health Institute (1R01AI173138-01A1) in the United States of America. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.
References
- 1.Antimicrobial Resistance Collaborators. Global mortality associated with 33 bacterial pathogens in 2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2023;400(10369):2221–48. doi: 10.1016/S0140-6736(22)02185-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bai A, Lo CKL, Komorowski A, et al. Staphylococcus aureus bacteremia mortality: A systematic review and meta-analysis. Clin Microbiol Infect. 2022;28:1076–84. doi: 10.1016/j.cmi.2022.03.015. [DOI] [PubMed] [Google Scholar]
- 3.Holland TL, Raad I, Boucher HW, et al. Effect of Algorithm-Based Therapy vs Usual Care on Clinical Success and Serious Adverse Events in Patients with Staphylococcal Bacteremia: A Randomized Clinical Trial. JAMA. 2018;320(12):1249–58. doi: 10.1001/jama.2018.13155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Westgeest AC, Buis DTP, Sigaloff KCE, et al. Global Differences in the Management of Staphylococcus aureus Bacteremia: No International Standard of Care. Clin Infect Dis. 2023;77(8):1092–101. doi: 10.1093/cid/ciad363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Baddour LM, Wilson WR, Bayer AS, et al. Infective Endocarditis in Adults: Diagnosis, Antimicrobial Therapy, and Management of Complications: A Scientific Statement for Healthcare Professionals From the American Heart Association. Circulation. 2015;132(15):1435–86. doi: 10.1161/CIR.0000000000000296. [DOI] [PubMed] [Google Scholar]
- 6.Delgado V, Ajmone Marsan M, de Waha S, et al. 2023 ESC Guidelines for the management of endocarditis: Developed by the task force on the management of endocarditis of the European Society of Cardiology (ESC) Endorsed by the European Association for Cardio-Thoracic Surgery (EACTS) and the European Association of Nuclear Medicine (EANM) Eur Heart J. 2023;44(39):3948–4042. [Google Scholar]
- 7.Miller WR, Seas C, Carvajal LP, et al. The Cefazolin Inoculum Effect Is Associated With Increased Mortality in Methicillin-Susceptible Staphylococcus aureus Bacteremia. Open Forum Infect Dis. 2018;5(6):ofy123. doi: 10.1093/ofid/ofy123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lee S, Kwon KT, Kim H-I, et al. Clinical implications of cefazolin inoculum effect and β-lactamase type on methicillin-susceptible Staphylococcus aureus bacteremia. Microb Drug Resist. 2014;20(6):568–74. doi: 10.1089/mdr.2013.0229. [DOI] [PubMed] [Google Scholar]
- 9.Prosty C, Noutsios D, Lee TC, et al. Cefazolin versus Antistaphylococcal Penicillins for the Treatment of Methicillin-Susceptible Staphylococcus aureus Bacteremia: A Systematic Review and Meta-Analysis. Clin Microbiol Infect. 2025:S1198-743X(25)00235-6. doi: 10.1016/j.cmi.2025.04.045. [DOI] [PubMed] [Google Scholar]
- 10.Tleyjeh IM, Kashour T, Mandrekar J, Petitti DB. Overlooked Shortcomings of Observational Studies of Interventions in Coronavirus Disease 2019: An Illustrated Review for the Clinician. Open Forum Infect Dis. 2021;8(8):ofab317. doi: 10.1093/ofid/ofab317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Tong SYC, Mora J, Bowen AC, et al. The Staphylococcus aureus Network Adaptive Platform Trial protocol: New tools for an old foe. Clin Infect Dis. 2022;75(11):2027–34. doi: 10.1093/cid/ciac476. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mahar RK, McGlothlin A, Dymock M, et al. A blueprint for a multi-disease, multi-domain Bayesian adaptive platform trial incorporating adult and paediatric subgroups: the Staphylococcus aureus Network Adaptive Platform trial. Trials. 2023;24(1):795. doi: 10.1186/s13063-023-07718-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Anpalagan K, Dotel R, MacFadden DR, et al. Does adjunctive clindamycin have a role in Staphylococcus aureus bacteremia? A protocol for the adjunctive treatment domain of the S. aureus Network Adaptive Platform (SNAP) randomized controlled trial. Clin Infect Dis. 2024;79(3):626–34. doi: 10.1093/cid/ciae289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.de Kretser D, Mora J, Bloomfield M, et al. Early oral antibiotic switch in Staphylococcus aureus bacteraemia: The Staphylococcus aureus Network Adaptive Platform (SNAP) Trial Early Oral Switch Protocol. Clin Infect Dis. 2024;79(4):871–87. doi: 10.1093/cid/ciad666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Liu C, Bayer A, Cosgrove SE, et al. Clinical practice guidelines by the Infectious Diseases Society of America for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children: executive summary. Clin Infect Dis. 2011;52(3):285–92. doi: 10.1093/cid/cir034. [DOI] [PubMed] [Google Scholar]
- 16.Henderson A, Cheng MP, Chew KL, et al. A multi-site, international laboratory study to assess the performance of penicillin susceptibility testing of Staphylococcus aureus. J Antimicrob Chemother. 2023;78(6):1499–504. doi: 10.1093/jac/dkad116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Legg A, Meagher N, Johnson SA, et al. Risk Factors for Nephrotoxicity in Methicillin-Resistant Staphylococcus aureus. Bacteraemia: A Post Hoc Analysis of the CAMERA2 Trial. Clin Drug Investig. 2023;43(1):23–33. doi: 10.1007/s40261-022-01204-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Baldwin DS, Levine BB, McCluskey RT, Gallo GR. Renal Failure and Interstitial Nephritis Due to Penicillin and Methicillin. N Engl J Med. 1968;279(23):1245–52. doi: 10.1056/NEJM196812052792302. [DOI] [PubMed] [Google Scholar]
- 19.Klein JO, Finland M. The New Penicillins. N Engl J Med. 1963;269(20):1074–82. doi: 10.1056/NEJM196311142692006. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.