Editor’s Spotlight/Take 5: Patients With Nonanaphylactic Penicillin Allergy Are Not at Increased Risk of Allergic Events After Receiving Prophylactic Preoperative Cefazolin Prior to Closed Fracture Repair

Gregory J. Della Rocca MD (left) and Boris Mraovic MD (right)This one surprised me. I had thought we settled the question of which antibiotic to give patients before orthopaedic surgery, even patients with penicillin allergies (other than, perhaps, those with documented true Type 1/immunoglobulin–E mediated anaphylactic reactions and rare exfoliative reactions like Stevens-Johnson syndrome and toxic epidermal necrolysis). After all, a well-substantiated clinical practice guideline [2] came out way back in 2013 giving us the answer: It's cefazolin. That guideline [2] contained more alphabet soup than a middle-school sick day. It derived from the combined efforts of the American Society of Health-System Pharmacists (ASHP), the Infectious Diseases Society of America (IDSA), the Surgical Infection Society (SIS), and the Society for Healthcare Epidemiology of America (SHEA). I remember when it came out. Our hospital, like many others, did a self-assessment. The team charged with this effort at our shop surveyed nearly 15,000 surgical procedures, about 1700 of which were performed on patients who reported having penicillin allergies. They discovered that only a small minority (16%) of patients with a documented penicillin allergy—ranging from things that don’t sound like “real” allergies, like GI upset, up to anaphylaxis—still received cefazolin, and none of them experienced an adverse reaction [4]. Fewer than 0.1% of those with no reported history of beta-lactam allergy experienced an allergic reaction. Based on those findings [4], our hospital took all sorts of steps to create systems-level approaches that ensured better practices for antibiotic stewardship, specifically to increase the proportion of patients with non-life-threatening penicillin allergies who would receive cefazolin. Other hospitals did, too [11]. Those programs—which may include steps like updating local guidelines and order sets, educating surgeons and anesthesiologists, and greasing the skids for allergy evaluation before a scheduled procedure in effort to “delabel” charts with erroneous penicillin allergies—generally have been effective [5, 9, 12]. Such programs are important, because it’s been shown time and again that if your chart says you’re allergic to penicillin, you’re more likely to get a surgical site infection, presumably because your doctors may not be following these well-established guidelines [1, 13], and the drugs they may choose instead of cefazolin just aren’t as good for surgical antimicrobial prophylaxis. So, back to my surprise. In this month’s Editor’s Spotlight article in Clinical Orthopaedics and Related Research® [3], a team from a leading trauma center, under the guidance of Gregory J. Della Rocca MD, looked back on more than 5000 patients treated surgically for closed fractures. Among those who had non-life-threatening penicillin allergies, only a minority—fewer than 40%—received cefazolin. That low percentage, more than 10 years after the publication of a robust clinical practice guideline [2] suggesting it should probably be much higher, is what surprised me. Less surprising (but validating and important nonetheless) was that the administration of cefazolin was found to not be associated with an increased risk of an intraoperative adverse event of any sort. Likewise, when they focused only on allergic events, there was no increased risk; allergic events occurred in the same proportion of patients regardless of whether they received cefazolin or an alternative antibiotic, and that percentage was 1%. This study [3] seems important to share for two reasons. First, if this group—which includes several thought-leaders who really know their stuff—was still avoiding cefazolin in most patients who reported nonsevere penicillin allergies as late as the end of 2019 (which is when the capture ended for their study), it’s likely many others still are. After all, we know that incorporating new evidence—even high-quality evidence—into practice is a process that often takes decades or generations, not months or years [7]. Second, I had seen that one of the coauthors on this paper [3] wrote a letter to the editor of JAMA Surgery [8] regarding a strong, well-done meta-analysis [10] published in 2021 on this topic. In that letter [8], he recommended against cefazolin because of concerns about the potential for severe allergic reactions. Now, 5 years later, it appears that coauthor has changed his mind, presumably in response to the good data that his research team analyzed here [3]. That second item especially appealed to me, and I think it should appeal to you, too. Notwithstanding the behavior of many of our elected leaders (on both sides), I think changing one’s mind in response to good-quality, new information is a sign of wisdom. Learning what tipped the scales for the authors of this month’s Editor’s Spotlight paper [3] can help each of us make good decisions on this topic as well as on the many others in the future when new discoveries call upon us to change our minds. Their study is worth reading whether or not you take a lot of calls or fix a lot of fractures, since the findings here almost certainly generalize well to other clean, elective orthopaedic interventions. Another reason this Take 5 interview is worth reading is that Dr. Della Rocca invited his coauthor, Boris Mraovic MD—who wrote that letter to JAMA Surgery over 5 years ago [8]—to join him in answering some questions, too. So let's go behind the discovery with two coauthors of “Patients With Nonanaphylactic Penicillin Allergy Are Not at Increased Risk of Allergic Events After Receiving Prophylactic Preoperative Cefazolin Prior to Closed Fracture Repair” [3]. Take 5 Interview with senior author Gregory J. Della Rocca MD and coauthor Boris Mraovic MD on “Patients With Nonanaphylactic Penicillin Allergy Are Not at Increased Risk of Allergic Events After Receiving Prophylactic Preoperative Cefazolin Prior to Closed Fracture Repair” Seth S. Leopold MD:Congratulations on this well-done study [3]. What’s changed vis-à-vis antibiotic selection at your institution based on what you’ve discovered here, and what steps have you taken to make sure those changes “stick”? In my experience, a lot of quality-improvement programs result in changes, but then people (and the systems in which they work) tend to slide back into old habits… Gregory J. Della Rocca MD: Thank you very much for this question. As of the moment I’m writing this (November 2025), nothing has really changed. Our personal practice is to administer cefazolin preoperatively to patients with fractures, even when they claim to have nonanaphylactic penicillin allergies, but it seems as though the institution-wide practice still involves administering clindamycin or vancomycin to patients who are allergic to penicillin. Like many other institutions, we have substantial turnover of staff, from pharmacy to operating room personnel to anaesthesia practitioners to orthopaedic trainees. Newer clinicians tend to have an almost knee-jerk reflex to give vancomycin or clindamycin to patients who are allergic to penicillin. When we catch the problem, we correct it, but a specific, institution-wide guideline would be most helpful. Since our institution does not have one, we are in conversations with the central pharmacy regarding the development of one. Part of what we hoped to convey in our manuscript [3] is that it is generally OK to administer cefazolin to a patient with a nonanaphylactic penicillin allergy. We suspected this would be the case, and prior guidelines and meta-analyses and primary research support this suspicion. The more evidence that can be produced in support of that statement, the more likely the statement will be accepted, because the general feeling among many practitioners still seems to be that cefazolin should be avoided in patients who are allergic to penicillin. Dr. Leopold:I was most intrigued by the letter to the editor in JAMA Surgery [8] that one of the coauthors on this paper (Dr. Mraovic) wrote about 5 years ago that made (more or less) the opposite recommendation to the one you (and he) make in the article in this month’s CORR [3]. And based on the practice pattern at your center during the span of your study, it’s clear enough that many other orthopaedic surgeons at your center felt as Dr. Mraovic did. This suggests some kind of journey took place, some minds changed there. Can you tell our readers a bit about how that happened? Boris Mraovic MD: Thank you for your interest in this important topic and how it has changed over time. The science of medicine is continuously evolving, and it should always be evidence-based. Even “solid” recommendations like practice guidelines can look like “alphabet soup” 10 years later, and they should be challenged and/or confirmed by new data. I believe anesthesiologists like me have different clinical experiences with cefazolin anaphylaxis compared with orthopaedic surgeons. We need to deal with crashing patients under anesthesia during surgery, while surgeons often deal with ambulatory patients who develop surgical site infections (SSI) postoperatively. Anesthesiologists are more likely to be “conservative” in avoiding administering cefazolin in patients who might have dual allergy—especially if they have dealt with severe intraoperative anaphylaxis—and they might ask for more details about “penicillin allergies,” which may or may not end up in the updated chart. This could explain the low percentage of cefazolin administered to our patients with non-life-threatening penicillin allergies. In my JAMA Surgery letter [8], I argued that data on true prevalence of cefazolin anaphylaxis is not known; the risk of anaphylaxis to cefazolin vs SSI is 1 episode of anaphylaxis for every 300 to 7000 SSIs. I concluded that until we have more data, we should take a “conservative” approach. If the incidences of SSI and cefazolin anaphylaxis are low, then more liberal cefazolin administration could be considered. I really liked our study [3] because it helped answer these fundamental questions. We did not find a difference in adverse reactions between cefazolin and secondary antibiotics. If this holds true, then cefazolin should be given to all patients with penicillin allergies, except those with type 1 anaphylactic reactions. Finding out what type of penicillin allergy a patient has could be a critical part of clinical practice. Sometimes, it is difficult to obtain history from patients, especially ones with fractures or after multiple trauma. However, this study shows that if they do not a have clear history of anaphylaxis, then cefazolin can be given safely. Dr. Leopold:I don’t want to seem like I’m picking on anyone, so let me share an instance in which I changed an everyday practice pattern and how hard it was. When I first started in practice, I did about one-third of my knee replacements as same-day bilateral procedures. I thought it was efficient and effective, but data emerged showing that by not staging these procedures a few months apart, I was exposing patients to excess risk of serious complications and perhaps death. I shared this story in an earlier Spotlight [6], so to briefly summarize, although I was well intentioned, it took some time to accept that I almost certainly was hurting people and needed to change my approach. But getting to that decision felt like passing through the five stages of grief. Could you comment on the emotional challenges of letting go of a practice that you thought was the right choice as new data suggested otherwise? What helped you (or those in the group who initially felt cefazolin was the wrong choice) feel better about making a change? Dr. Mraovic: That is a great example. Unfortunately, on more than one occasion, I’ve heard, “I have been doing it like this for 20 to -30 years,” which I interpret as, “Who are you to tell me to change my practice?” Old habits die hard. As a clinician and researcher, I always look for the best evidence and try to put my patients’ safety first. In my opinion, we should all actively seek to improve current practice. More importantly, we must be open to change our practices as new data emerge and not just follow “dogmas” we once learned. For me, personally, changing practice as new evidence emerges feels like moving forward. It feels like I am a better clinician, and that I am serving my patients better. Sometimes, as you mentioned, learning about changes in best practice recommendations may elicit “grief,” but it may be helpful to remember that these practice changes ultimately benefit the patients—in this case, helping them avoid a postoperative infection. It’s also helpful to think that we might have avoided these complications previously if only the evidence had been present before. Ultimately, we as clinicians should be humble and open-minded, and should do what is best for our patients. Dr. Leopold:You had relatively few allergic events, so the 95% confidence intervals around your no-difference finding about allergic reactions were wide [3]. A skeptic might say that you were underpowered, but I’m not so worried. To my eye, the near-symmetry of the 95% CI around the odds ratio for allergic reactions (from about one-fifth to 6x), as well as your conclusion’s concordance with other studies done in many surgical specialties [10] (and with the leading guideline on this point [2]), gives me confidence that your recommendation favoring cefazolin in patients who report a nonanaphlactic allergy is the right one. Yet you still call for “prospective cohort studies” on this topic, which, as you know, are not easy to do. Sometimes, I think we have to declare a topic settled so we can use scarce research resources to explore topics where controversy persists. Why do you think this one is still, at least to some degree, up for debate? Dr. Della Rocca: Great question. I think about this problem of underpowered studies quite often when reviewing the evidence that supports orthopaedic treatment decisions. Some orthopaedic conditions are too uncommon to yield to large prospective trials about their treatment. Some other questions are just not likely to be funded. As you are aware, adequately powered, prospective, randomized trials are expensive and time-consuming. Our question about the safety of cefazolin administration to patients who are allergic to penicillin (instead of vancomycin or clindamycin) is unlikely to be successful in securing funding for a large prospective study. At risk of being cynical, this may have something to do with the fact that the drugs in question are all available as generic formulations, and so the pharmaceutical industry is unlikely to fund such a prospective trial. And we are all aware of how difficult it is for excellent research, in orthopaedic surgery and throughout the field of medicine, to secure public funding. I agree with your sentiment about using scarce research resources to explore topics where controversy persists. The most likely explanation for continued debate here is a lack of knowledge. An institution-wide guideline would go a long way to helping with this problem. Many practitioners will continue practicing as they have always done, even if the evidence is 20+ years out of date. Obviously harmful practices are quickly discarded, but practices where harm is either minimal or not easily obseved (for example, an SSI that happens after administration of a secondline antibiotic for preoperative prophylaxis in a patient allergic to penicillin and is found during postoperative follow-up isn’t seen by the anaesthesia professional, as Dr. Mraovic mentioned) may persist. Constant movement of healthcare professionals between institutions may result in the return of old habits. Dr. Leopold:What about patients with severe penicillin allergies like anaphylaxis—does your center try to get them seen by an allergist, as some have recommended [5], or do you just use another antibiotic? If the former, how successful are you at making this happen? My observation is that it can be pretty hard to do in the “real world”… Dr. Della Rocca: We do not currently have patients who claim to have a severe penicillin allergy see an allergist. Our practice is a trauma practice, so their fracture repairs are happening on a relatively rapid basis after their injuries occur. I do not know if the practice of our colleagues who perform elective orthopaedic surgery is to have those patients evaluated by allergists, but I suspect not. Our institution does not have a dedicated allergist; instead, the allergy clinic is overseen by two colleagues, one from otolaryngology and one from dermatology. The wait time for a new patient to be evaluated is between 1 and 2 months. Aside from Jones et al. [5], we are not aware of many recommendations of this type in orthopaedics. The general tendency in practice, therefore, is to use another antibiotic. Our joint reconstruction partners will generally give prophylactic vancomycin and aztreonam, instead of cefazolin, to patients who have histories of severe penicillin allergies before elective joint arthroplasty. Your “real world” observation hews close to the mark.