Dear Editor,

We read with great interest the article titled “Impact of silencing automated penicillin cross-reactivity alerts on perioperative antibiotic prescribing and surgical site infection rates” by Durkin et al. ^{Reference Durkin, Nordman and Bewley1}

The authors examined the impact of silencing automated penicillin cross-reactivity alerts on antibiotic prescribing and surgical site infection (SSI) rates in the perioperative setting. ^{Reference Durkin, Nordman and Bewley1} Using interrupted time series analysis across six hospitals, the study focused on patients (n = 6,204) undergoing CABG, colorectal surgery, hip replacement, and hysterectomy between March 2020 and February 2022. Approximately 15% had a documented penicillin allergy—438 pre-intervention and 486 post-intervention. Following the alert change, cefazolin use among allergy-labeled patients increased modestly from 33% to 39%, and all beta lactam use increased from 71.5% to 78.4%. SSI rates remained statistically unchanged: 2.9% vs 4.0% (P = .32) in allergy-labeled patients and 3.1% vs 3.3% (P = .99) in controls. Notably, 45% of SSIs were intra-abdominal. The authors concluded that multifaceted strategies are needed to improve outcomes, particularly given that single implementation strategies do not consistently lead to meaningful practice change. ^{Reference Rogal, Yakovchenko and Waltz2}

We note that the study may have been underpowered to detect meaningful changes in SSI rates. While the authors acknowledge this limitation, its implications are not fully explored. A 6% increase in cefazolin use equates to roughly 29 additional patients receiving the antibiotic. Evidence supporting cefazolin’s effectiveness in reducing SSIs, particularly in clean-contaminated procedures, requires a number needed to treat (NNT) of 20–50, with most estimates clustering around 30–40. ^{Reference Pop-Vicas, Johnson and Safdar3,Reference Seidelman, Mantyh and Anderson4} Previous work by Blumenthal has shown that the NNT needed to reduce an SSI when cefazolin is compared to non-beta lactam antibiotics, in patients with reported allergy, is around 112 to 124. ^{Reference Blumenthal, Ryan, Li, Lee, Kuhlen and Shenoy5} Using the numbers of SSIs from Blumenthal et al through PS Power with 80% power, you would need 11,462 patients with penicillin allergies to meet the power. ^{Reference Dupont and Plummer6} Therefore, the modest increase in cefazolin use reported in this study is unlikely to allow for a robust assessment of changes in SSI rates.

The non-cefazolin beta-lactams used in the study were not described. As some beta lactams, such as cefoxitin and cefotetan, have higher SSI rates than cefazolin, it is difficult to benchmark the baseline rate of SSIs with 37–40% of prophylaxis use being beta lactams other than cefazolin. ^{Reference Pop-Vicas, Johnson and Safdar3} Hence, the high use of other beta lactams in this population may have impacted the uptake of cefazolin with the intervention.

Furthermore, the benefit is greater in higher-risk surgeries and when cefazolin is administered within 60 minutes before incision and dosed appropriately by weight. ^{Reference Calderwood, Anderson and Bratzler7} However, in this study, the median administration time exceeded one hour, with a median of 82 minutes (IQR 51–108).

Finally, we agree with the authors that SSI rates are influenced by patient factors, procedural risks, and adherence to infection control practices. Without adjusting for these confounders, conclusions regarding SSI outcomes should be interpreted with caution. It is also important to note that the intervention described in this study is safe, ^{Reference Macy, McCormick and Adams8} which supports its consideration in clinical practice despite the need for further evaluation of its impact on SSI rates.