Is it safe to administer cefazolin in penicillin-allergic patients?

Welcome back to Depth of Anesthesia. This is a podcast where we critically explore our clinical practices. I'm David Hao and I'm an anesthesia resident at the Massachusetts General Hospital. This is a special day today because for the first time in our podcast history, we have a guest who is not named Dr. Daniel Stowy Kinevka Today I'm super excited to welcome Dr. Matt Vanneman to the show. He's an attendant cardiothoracic anesthesiologist and was chief resident at our program.

Matt, welcome to the show. Very honored to be here, David. Glad to help you break your barriers and set new uh records. This is amazing. So we'll start with a case today.

And our case is that of a 67-year-old woman with a history of left knee osteoarthritis who is presenting for a left total knee replacement After induction, the resident recalls that the patient had reported a penicillin allergy, and instead of administering cefazolin, also known as ANCF, the resident selects vancomycin. That is definitely something I see very commonly. And I remember even back when I was a resident that the common teaching that I always heard from many of my attendings was If a patient has a penicillin allergy, there is a chance that the patient may have a cross-reactive allergy to cefazolin, and then as a result, we should administer vancomycin. To help prevent the risk of that patient from having an allergic reaction to cefazolin.

So rather than administer the standard or cefazolin, we instead chose to administer a betalactam alternative, in this case being vancomycin. Yeah, and I've definitely seen this practice habit before as well. The claim here being that in penicillin allergic patients, a non-beta lactam antibiotic should be administered. So for our listeners, what's the level of belief that you have in this claim? And what do you think is the level of evidence that supports this claim?

Do you think that there's no evidence but stands to reason? Or do you think that there's expert consensus or maybe mixed trials? Or is it all correlation or is there a chance that we even have high level causal evidence or randomized controlled trials? And Matt, before we get started, I think it would be helpful for us to break down this claim a little bit further in terms of exploring some of the rationales that people give when they go ahead and support this claim. So I think about this kind of in two uh separate claims within the first claim.

So the overall claim is we shouldn't give patients with a penicillin allergy uh a cefazolin because there's a chance they may have a dangerous allergic reaction and also the the alternative antibiotic that we're going to choose will be just as good as cefazolin So we shouldn't take the risk that would be inherent in administering cephalin to a penicillin allergic patient. So there are two claims here 1. An antibiotic that is an alternative to cefazolin, but is not cefazolin will be equally efficacious at preventing a surgical site infection when compared to cefazolin 2. Administering cefazolin to a patient with a known penicillin allergy places that patient at risk of a possibly dangerous uh allergic reaction as a result of receiving cefazolin And therefore we should not administer cepizolin because it is actually not safe for patients with a penicillin allergy.

That's kind of how I think about this claim into kind of two separate parts, and I think we should explore each part separately. So let's go ahead and uh start investigating the first portion of that claim. And I know we have two articles which are gonna tackle different components. And the first is looking at the impact of a reported penicillin allergy on the general population. And then we'll dive into what the impact of a reported penicillin allergy is. on surgical site infection risk, which is very relevant to us as anesthesiologists.

And the first paper that we're going to be talking about, I believe, is by Dr. Blumenthal, who's actually one of our colleagues here at the MGH. And this is about uh the impact of a documented penicillin allergy in the general population. Tickets Started, uh, there's this study by uh Dr. Blumenthal he had mentioned, and this is a study of about 300,000 people uh who are in the United Kingdom in general practice from 1995 to 2015. And what Dr. Blumenthal and her colleagues did was they took 64,000 of these 300,000 patients with a known penicillin allergy. and then match them based upon age, comorbidities, and other factors with 240,000 other patients who did not have a penicillin allergy So it's a matched cohort study.

On the one hand, 65,000 patients with a penicillin allergy. On the other hand, 240,000 patients without a penicillin allergy, matched again on comorbidities and other factors. So what they've hypothesized was that patients with a penicillin allergy were more likely to receive non-beta-lactam alternative antibiotics things such as vancomycin or clindamycin, and they hypothesize that because of this they there would be an association of increased risk of infection with resistant organisms such as methacill-resistant staphorius or MERSA for short, or uh clostridium defeceile, which is a result of destruction of normal flora, most frequently associated uh in our minds at least with things like claindamycin, and then that resulting in uh clostridium defecil So one of the difficult things with retrospective studies is the whole concept of assigning causality.

So did the authors here mention that or or discuss that at all? Yeah, excellent question, David. So to start with, what the authors did was they compared the two groups, the non-penicillinallergic group with the penicillin allergic group. And the first thing they looked at is what antibiotics did each of these groups receive? And so the first thing they saw was, unsurprisingly, patients who were penicillin allergic were much more likely to receive non-beta-lactam antibiotics.

So, for example, they found that patients who were panasallergic were fourfold more likely to receive macrolides, fourfold more likely to receive clindamycin, and twofold more likely to receive fluoroquinolides So again, this probably isn't a shocking finding. They have a known penicillin allergy. There's been discussion about danger of beta-lactam antibiotics in patients with a penicillin allergy, and therefore they should get alternatives. What was interesting is that they found that patients with a penicillin allergy were also more likely to develop multi-drug resistance infections. So For example, they were about 70% more likely to uh have an infection with methacillin resistant, Staphoreus, or MERSA, and approximately 25% more likely to have an infection with Clostridium Dfeecil.

So this isn't necessarily intuitive. There's no obvious reason why penicillin allergies should be directly linked to uh infection with these multi-drug resistant organisms So a lot of what you're describing seems more like correlation as a positive as opposed to causation. So is there any evidence in this paper to suggest that there might actually be a causality component of the data Yeah, so that's obviously very important to tease out. So what the authors in this study did was they took a very a thorough multivariate logistic regression to assign as much as they could how much uh correlation was assigned to each separate piece of of the puzzle. So, you know, if someone was a certain age, maybe that put them at a higher risk of MRSA, if they had certain comorbidities, that had a different correlation to MRSA, or if they were hospitalized or something.

So they kind of went through a very long list of possible covariates and confounders to determine which factors were more likely to be associated with the infection, with these multidrug-resistant infections. And the one of the factors they looked at in this regression was the receipt of non-beta-lactam alternative antibiotics and the increased risk of the having a multidrug infection. And what they found was that the beta-lactam alternative antibiotic was associated with a fifty five percent increase in the risk of uh developing a MERSA infection, and a thirty five percent risk of increased risk of developing a CDF infection that could likely be attributed to receiving these alternative antibiotics. Gotcha. So to summarize this paper, it seems like three hundred thousand patients, retrospective cohort We have correlation based on what the auditors are saying that a known penicillin allergy increases your risk of serious infections, including MERSA and C. diff.

Yeah, it certainly seems that way, but again, it's not a randomized controlled trial. So we aren't randomizing people to receive either cefazolin or clindamycin and then seeing who gets more multidrug-resistant infection It's just not what we're doing. So it's hard to say with certainty that there is a causal effect between receiving, for instance, clindamycin and then developing uh MRSA at a higher frequency than if they had received cephozolin. Nonetheless, it certainly suggested that there may be some relationship between receiving a non-beta-lactam antibiotic and subsequent development of a multidrug-resistant infection in the general population cohort that was studied in this paper. Gotcha.

So moving on now, Blumenthal has a second paper, correct? And this is the impact on surgical site infection risk Yeah, this is a very nice study, I think. So what Blumenthal and her colleagues did was in a retrospective cohort study looking at all of our own Massachusetts General Hospital data for a five-year period between 2010 and 2014. They looked at all patients undergoing a kind of a broad swath of surgical procedures. So these are patients undergoing hip orthroplasty knee arthroplasty, hip serectomy, a colon surgery, or coronary artery bypass crafting.

So really kind of surgery of across all parts of the body with many different uh types of surgery represented And they took this cohort of approximately 8,400 patients and then matched them and divided them basically if a patient had a penicillin allergy. or if they did not have a penicillin allergy and match them according to all the other standard kind of confounders and covariates that we could think of. And then the outcome that they were looking at was the risk of developing a surgical side infection postoperatively after uh completion of their operation. and seeing uh if having a penicillin allergy uh was correlated again with uh the risk of developing a specific surgical side infection And basically what they found was that uh if patients had a known penicillin allergy, this substantially increased their risk of developing a surgical site infection by approximately 50%.

So patients with a penicillin allergy had a 50% increased risk of developing a surgical side infection compared to patients who did not have a penicillin allergy. Now was this risk because they were getting alternative antibiotics? Again, very good question. So it's hard to again parse causality in these type of cohort studies. Nonetheless, uh as part of the logistic regression, they uh try to determine and assign uh kind of a relative risk of receiving an alternative non-beta -lactam antibiotic and seeing how that was associated with developing a surgical side infection So in this uh cohort, uh, if patients with a penicillin allergy were far more likely to receive alternative antibiotics rather than cefazolin.

So for instance, um they say if you had a paniculum allergy, uh you had a 49% chance of receiving clindamycin, uh 35% chance of receiving vancomycin And a 24% chance of receiving gentamycin, and only a 3% chance of receiving cephalin. So if you had a penicillin allergy, it was extremely likely you would receive a non-beta -lactam alternative antibiotic. So what Blumenthal and her team did was they, in their logistic regression, uh tr accounted for this alternative antibiotic administration and tried to determine how much of the alternative antibiotic accounted for the increased risk of surgical inside infection compared to all the other factors that may be related. And what they found was uh that in fact the alternative antibiotic uh completely explained the increased risk of surgical side infection in this population cohort that she was looking at suggesting that at the very least these alternative antibiotics were may not be as good as cefazolin at preventing surgical site infections Okay, so now both of these studies seem to suggest that having a reported penicillin allergy, at least from a retrospective look and assigning correlation, it's not great.

Are there any other papers in the perioperative period that would suggest otherwise? Yeah, so good question. So there is another study in the orthopedics literature. This is done by Dr. Tan and his colleagues And what the these group of orthopedics did was they did another retrospective cohort study of 10,000 patients. So it's a very large cohort.

They again divided them into patients who had either a penicillin or cephalosporin allergy. So in these this group is slightly different than the Blumenthal cohort, which was only penicillin allergy. So they divided them into patients with a penicillin or cephalosporin allergy and versus patients who had an al neither of those allergies. and then compared the risk of developing a surgical site infection between these two groups, again controlling for common confounders or covariants. And what they found was, uh unsurprisingly, patients who had a penicillin or cephalos-forein allergy were far more likely to instead receive vancomycin, an alternative antibiotic. um compared to patients who did not have those either of those allergies who were much more likely to receive cephalin.

And they found that patients who received vancomycin actually had no increased risk of a deep surgical site infection compared to patients who received cephozolin in their cohort. Now again this is an orthopedic study that was specifically focused on patients receiving total joint orthoplasty. So uh as a comparator to the Blumenthal study, uh the Blumenthal cohort was talking kind of a very broad set of surgical patients. So They included not only joint arthroplasty but also hysterectomy, colon surgery, coronary bypass graphic. Right.

So a variety of different pet patients. This cohort specifically looked at patients only receiving joint orthoplasty. But again, they did find that vencomycin was no different in preventing surgical side infection compared to cefazolin in their cohort So it it's an interesting finding that certainly is conflicting with the Blumenthal data. Gotcha. So what are your thoughts about these three studies?

So the one advantage of the Blumenthal data compared to the orthopedic study, the Dr. Stan study, is that it is a much broader surgical cohort compared to just a joint orthoplasty only. So in the joint orthoplasty in Dr. Tan's study, their risk of surgical site infection was only about 2%, and they were specifically looking at deep surgical site infections. Whereas in uh Dr. Blumenthal's this definition of surgical site infection was broader, so it comes more different infections. And most uh more of the infections actually from Dr. Blumenthal's study, actually most were coming from colorectal surgery or colon surgery. So this may not be completely surprising because vancomycin has excellent gram-positive coverage and very poor gram-negative coverage And actually, there interestingly, in Dr. Tan's study, they performed pre-specified subgroup analysis and found that patients who received vencomycin their surgical site infection was much more likely to be from a gram negative bacterium as opposed to a gram positive bacterium.

So in Dr. Blumenthal 's cohort, where most of the infections are from colorectal surgery, which is obviously has a large component of gram negative, uh That may be partially explained why the patients in her cohort had more surgical sign infections when they received alternative non-beta-lactam antibiotics. Another difference between the study is that patients in the TAN study only received vancomycin as a non-beta-lactam alternative, whereas in Dr. Blumenthal's study there was kind of this kind of amalgamation grab bag So it may be that vancomycin specifically is a superior alternative compared to clindamycin and gentamycin , but it's hard to say. We just don't have enough data to make those claims. To me, overall, it certainly suggests that there may be some evidence that the alternatives aren't as good as cephozolin.

That's a great explanation. I think that really captures the essence of those three studies. Now moving on to the second portion of the claim, and this is the portion where some of the rationale given includes concern for severe reactions to cephalosporins in penicillin allergic patients, and I think in that in part derives from the concept of cross reactivity, which we haven't talked about yet. Do we have any data for that portion of this claim? Yeah, so I to kind of dig into this a little bit further, I think we should really go back to the beginning and the origins of that claim so we can understand better where it comes from and then kind of understand how it matters to today.

So um I'll just talk about a brief review from 1978 in the Journal of Infectious Diseases by Dr. Petts. And where he was reviewing the claim of why patients who have penicillin allergies may have a cross-reactive allergy to cephalosporins. So the gist of the claim is that penicillin and cephalosporins are both built on the same beta-lactam backbone. And on top of that, penicillin has separate R chains, which are different chemical subgroups that come off the beta-lactam ring And may have antibodies directed against them. And those antibodies may mediate some type of IgE allergic reaction Similarly, if patients have those IgE antibodies to the penicillin, either the beta-lactam ring or the R subgroups off the ring, They may also have antibodies that are against the similar structure of the cephalosporins.

The thing that distinguishes penicillin from cephal sporins are these side chains off the common background So uh Dr. Petts in his review talked about a number of studies from the 60s and 70s looking at this possibility. So one of the studies that he refers to is that they would take patients who had received penicillin and had a known documented penicillin allergy, and they would take their blood and then perform a number of in vitro hemoglutin assays using in vitro cephalosporin preparations. So in one of the studies they looked at um cephalothin, which is a first generation cephalosporin, which I I've personally never used. Oh me neither. Uh and they basically said, if if you had an allergy to penicillin, they take your blood, combine it with cephalothin, and then uh see if there was a hemoglobin reaction, which would suggest the presence of anti-cephalothin antibodies And what they found was that patients who had a pancellin allergy were far more likely to have these antibodies present in their bloodstream.

So that was one kind of supportive in vitro evidence that the that there was this type of cross-reactivity. There was a second study which he cites which um basically did a similar type of study. So they had patients again allergic to penicillin, they took out their blood, and this time they looked at uh differing leukocyte stimulation tests and lymphocyte stimulation tests. And they saw if that a patient was allergic to penicillin, their blood was much more likely to stimulate a leukocyte response or lymphocyte response um to cephalothin compared to patients who were not allergic to penicillin. So again, they extrapolated from these data that patients with a penicillin allergy have are both Uh antibody serologies and other components that would activate the immune system when exposed to cephalothin, which they then interpreted and extrapolated to mean all cephalosporins.

And that's kind of where the basis of the this claim was made that patients with penicillin allergy uh in fact also may have cross-reactive responses to uh cephalosporins. Gotcha. So this is probably in part where some of the genesis of the concept of creat cross-reactivity came from. Exactly. And so again, obviously to state there are some flaws with the way of thinking about this type.

So first is this is all in vitro data. So um it's hard to know what that means when placed in in vivo peat patients and how that will work. Uh additionally, uh cephalothin, like I mentioned, is a drug that I've never personally given. It sounds like you have me given it. And in talking to my colleagues, I've never even heard of it being administered in our operating rooms.

Um, it may be used at other institutions So uh uh and we'll get into this in a second, um but the cephalothin will have specific R chains for that drug Whereas other cephalosporons, I have different R chain. Every cephalosporin has two different areas with sidechains, the R1 side chain and the R2 side chain. Uh if you look at say cephozolin, all of those sidechains are completely different. The R1 sidechain is different and the R2 sidechain is different. So what's important to think about is Are these sidechains similar, and that may cause a cross-reactivity to the antibodies in the bloodstream of the patient with the penicillin allergy?

Do we have any contemporary data to evaluate the safety of cephalosporins in penicillin allergic patients? Yes, there's a lot of data about this topic actually. Uh I think we can just talk about a few of the studies. They'll eventually get into a theme you'll see pretty quickly, um and then we can kind of go into possible some methodologic reasons why that might be the case. So to start with will be a study by uh Dr. Beltrott and colleagues at Nationwide Children's Hospital in Ohio.

And what they did was they uh also put together a cohort of uh patients uh who received who are penicillin allergic, so 513 patients who had a known penicillin allergy undergoing uh surgery. And so these patients uh underwent 624 different surgeries And what they found was actually that cephalosporins were administered in about one quarter of these cases to patients who had a known documented penicillin allergy. So you're saying that these patients definitely had the penicillin allergy documented and then they got the cephalosporents on top of that? Correct. Okay So um and turns out that uh cephalin was used in over 80% of these patients.

So of the patients with penicillin allergy, a quarter got cephalosporins, and of those got cephalosporins Over eighty percent got cefazolin. When tha that's the most common anbiocu we least use here at our institution for surgical side infection prophylaxis. Definitely And so what they found was that uh of those patients, uh there were no major uh reactions to as a result of receiving the cephalosporin antibiotic. There was one minor non-anaphylactic reaction as a result of receiving cephalozolin. But again, none of the patients had anaphylactic shock. none developed respiratory distress, and no one had really any major clinical outcome change as a result of receiving cephal cephazol, excuse me.

How was this study set up again? So again, this is a cohort study. So it's a bunch of patients who had a known penicillin allergy and they just looked at what antibiotics did they receive So a quarter of them received cephalosporin, and then the other three-quarters received an alternative, and most likely that was clindamycin. And interestingly, actually, what they found was that patients who received clintamycin, they had a reported adverse uh reaction event of 1.5%. which was actually higher than the patients who received cephalin or cephalosporins. They looked into these adverse reactions.

They found that six patients developed a rash clindamycin, two had uricaria, and and one had erythema. So again, this just at least suggests that uh clindamycin or alternative beta-lactin antibiotics may not necessarily be safer than cefazolin in patients with a documented penicillin allergy. Now the question I have for you about this study is What about anaphylaxis? Because I imagine there's a certain caution in interpreting this type of data if the whoever was administering these medications knew beforehand that the patient had a severe reaction and subsequently made decisions based on that. Yeah.

Exactly, and an excellent point. So one of the problems of this study is it's because it's a retrospective study, um it just is what did people do and looks at comparisons So it is likely, although it's not specified in the paper, that patients who reported a severe reaction or anaphylaxis to penicillin were less likely to receive stepzolin. Or put another way, the patients who actually receive cefazolin in this cohort are likely to have only reported either a minor reaction or a distant reaction. or reaction that wasn't concerning to the uh anesthetic provider such that they felt safe administering cefazolin to them. Uh as an example, I doubt that any of the patients who received cefazolin said to their NSC provider, Oh, I just got out of the hospital for anaphylactic shock from penicillin.

We just don't know that. So it's hard to say if that in patients who have a known severe penicillin allergy, if cefazolin is safe based upon this study. It's just we can't say in one way or another. Gotcha. And and again I think that speaks to the caution we have to have in drawing conclusions from retrospective data.

Exactly. I completely agree. So I know you have another s retrospective cohort study. This was by Goodman, right? And this was in a broader population of orthopedic patients Yeah, so this is a very nice study.

So what Goodman and her colleagues did was they looked at about 3,000 anesthetics delivered to adult patients receiving orthopedic surgery And what they looked at and identified approximately 400 patients who had a documented penicillin allergy and received cephalin during their procedure. What they found was, or they were investigating, was did any of these patients who had a penicillin allergy and got cephozolin, did any of them have any adverse reaction? And what they found was that of the 400 patients with a known penicillin allergy who received cephalozolin, uh only one had a possible allergic reaction And basically, what they did was they inferred the top possibility of an allergic reaction because this patient's anesthetic record was noted to have been administered diphenhydramine and hydrocortisone towards the beginning of the case However, in that one specific patient, there was no mention of any symptoms of allergic reaction.

There was no mention of respiratory distress, no mention of itching, no rash, no hives So it's unclear if the administration of the corticosteroid and the antihistamine in this case was related to cefosolin or some other uh factor Nonetheless, the thing that was most interesting was that of the other four hundred and twelve patients who had a known penicillin allergy Uh none of them, uh other than this single reaction, had any other type of reaction when they were administered a cephalosporin, which most commonly was cefazolin. So would you agree that the same limitations apply in the sense that this is a retrospective study and there's a chance that the clinician felt like the patient had a true anaphylactic reaction? then they would make subsequent management steps based on that knowledge.

And so the actual group that were encompassing are those who are likely to not have a true anaphylactic reaction report it. Yeah, correct. I think that's exactly right. So again, this study looked at 3,000 patients who had a panicillin allergy. So the question was, of those patients, the 413 who received a cephalosporin I don't know, they probably didn't have a severe allergy because uh they had reported severe allergy, it's likely that the anesthesia provider would not have administered cezolin.

Uh we can't say for sure one way or another. It's hard to know. They don't report these data in this particular paper, but at least in the patients who did receive cephozolin, uh there was not a significant instance of an adverse reaction So those are the two studies that we're we took a look at in the surgical population. I know you have one more and this is in a much broader group and it's a larger cohort. So, our third study T that we're looking at is a very large retrospective cohort study out of Kaiser in California And what they ended up doing was they ended up looking at all of their plan members over roughly a three-year period and then ask the question of what happens when their plan members receive cephalosporin antibiotics.

And so they compared plan members who had no history of any type of drug allergy versus plan members who had a known history of a penicillin allergy. And then they looked retrospectively to see how these patients did after receiving a dose of either an oral or an IV cephalosporin. So this isn't the surgical population, but it's kind of all comers. All people who come, either they have no history of drug allergies or history of a penicillin allergy, they then receive a cephalosporin at some point in their clinical care, and what happened? So what they did and they ended up finding was that patients who had no history whatsoever of receiving encephalosporin 30 days later had approximately a 0.4 % chance of having a new cephalosporin allergy documented Meaning that after they received the cephalosporin allergy the cephylsporin antibiotic, excuse me, um there was a 0.4 % chance 30 days within 30 days a new cephalosporin allergy was added to their electronic milk record Which they said suggested that they may have had some type of allergic or adverse reaction.

Now, in comparison, patients with a known penicillin allergy had approximately a Three to four fold increased risk of having a new cephalosporin allergy added to their allergy list after receiving a cephalosporin antibiotic. So what they do is again if they got an antibiotic cephalosporin antibiotic So within 30 days, a new cephalosporin allergy was added to their allergy list at a fourfold increased risk. But what that actually means is that of the patients with a known penicillin allergy Approximately 1.2% received a new cephalosporin allergy added to their electronic milk record after receiving a cephalosporin. So overall the rate was quite low, meaning approximately 99% of patients who received the cephalosporin and were known to be penicillin allergic did not have a cephalosporin uh allergy added And so I just want to clarify the way that the study determined whether or not something was an adverse reaction was through the electronic medical record.

So if a adverse reaction or allergic reaction was listed within 30 days of receiving that antibiotic. That's correct. So if the patient called up and said, Oh, I now have a new rash and someone added um the allergy to the new cephalosporin with a reaction of rash to the electronic milk record That is what the they were quantifying. And again, in the no allergy population, that only happened approximately 0.4 % of the time, whereas patients with a known penicillin allergy, that happened approximately 1.2% of the time. Triple the you know triple or three or four times more, however, overall a low rate.

The best part though I thought about the study was The uh the question asked was, how frequently was anaphylaxis documented into the uh medical record? Because that would be a fairly important finding if we knew that cephalosporins were giving anyone with a penicillin allergy anaphylaxis. That would be very important to know. And so what they found was actually the overall rate of anaphylaxis was extremely low. They found that of the approximately 260,000 patients with a known penicillin allergy who received a cephalosporin, only three had documented instances of anaphylaxis after receiving their cephalosporin.

Now, in comparison, in patients who had no history of any drug allergies, there were seven documented instances of anaphylaxis. in patients who had no history of any type of drug allergy out of a total patient co-population of 3.2 million So 7 out of 3.2 million are compared to 3 out of approximately 275,000. So statistically speaking, because the overall rate was so low, there was no statistical difference in the rate of anaphylaxis. in this cohort who had a penicillin allergy or who had no history of drug allergies who then received cephalosporin antibiotics. Now with this data, as usual, we're limited by the same limitations that we've had with the other two studies in that this data is retrospective.

So do we have any prospective data on the incidence of reaction in penicillin allergic patients who received a cephalosporin? Yeah, good question. Uh it's a very tricky question to get at because uh if we thought that patients who had a penicillin allergy would be in very high danger if we were to give them the cephalosporin. It's hard to prospectively say, oh, we're going to try this out to see if this uh medicine we give you gives you anaphylaxis. But what we do have is kind of a combined set of 12 studies that were put together by Blumenthal.

And basically what this did was they took 400 patients who have been pre-identified as having a penicillin allergy by a known skin test. So this isn't like a type of uh GI upset reaction. We know this is a type 1 hypersensitivity reaction from these patients. And what he he uh what Blumenthal did was uh she looked at what happened to these patients when they received the cell philosporin for other reasons. And again, this is a pooled set of 400 patients across 12 studies.

And what they found was of these 400 plus patients, only eight had a reaction to the cephalosporin antibiotic that they received. So the overall rate of reactions in this cohort was approximately 2%. And the reason this information is important is that this is in patients who would probably be at the highest risk of having a cross-reactive reaction Because these are patients we know have type 1 hypersensitivity reactions from allergy skin testing. However, that being said, there is a small drawback to the study, which is that penicillin allergy skin testing is not perfect. So the positive predictive value of penicillin allergy skin testing varies uh from 40 to 100 percent per this article by Dr. Blumenthal.

And so what that means is in the original cohort of 417 patients If we assume that the skin testing was perfect, had a positive predictive value of 100%, meaning that if all of all the patients who got the skin test that tested positive All of them in fact are allergic to penicillin. If we assume that to be the case, the one the perfect case of 100% positive predictive value of the skin testing. Then we would uh justifiably conclude, okay, eight of those 417 patients had a reaction, therefore the cross-reaction rate would be roughly in the 2% range. We'd say, okay Sweet. That's awesome.

The downside though is let's assume a more uh less optimistic scenario. Let's let's assume that the positive predictive value of penicillin skin testing is only 50% So that means that if I test two people with uh this penicillin skin test, and both of them test positive That means if the positive predicted value is only 50%, it is likely that in fact only one of them has a true type 1 hypersensitivity reaction to penicillin. So the reason this matters is because if we assume that the penicillin skin test only has a positive predictive value of 50%, that means of the roughly 400 patients we started with who tested posit by skin testing for a penicillin allergy, Only half of those patients would have a true type 1 hypersensitivity reaction.

The other half would just be a false positive result from the erroneous testing. Which would mean that of our cohort, only 200 have a true type 1 hypersensitivity reaction, and of those 200 Eight of those had reactions to cephalosporins, meaning that the adverse reaction rate to cephalosporins in a truly allergic penicillin population could be as high as four percent And that would certainly exceed the frequency that we saw in Dr. Macy's paper, which was closer to 1%. Uh and again, even though this entire frequency cannot be totally attributed to the concept of cross-reactivity, it's definitely not ins inconsequential, right? Exactly. And so uh uh one thing I do just want to bring up is here, so we're talking about up to a two to four fold increased risk of having an allergic reaction or an adverse reaction to cephalosporins in the patients who are labeled as penicillin allergic.

But in an absolute risk scale, the difference is going from 1% to at most 4%, somewhere between 2 and 4%, depending upon how we think about it. Uh so the absolute risk difference is overall low. And the other important point is that these are just all allergic reactions combined This isn't talking about the most severe ones, which we would probably be most concerned about giving our patients. So things like anaphylaxis, which were talked about in the Kaiser paper by Dr. Macy, those uh again were overall rare events, which are just something important to remember. One of the most interesting parts about this data to me is that the ranges that we're talking about seem to be at odds with the 8 to 10% range that I see commonly quoted.

Do we know where that 8 to 10% range comes from? Is that from data or is that, you know, made up? Yeah, so I actually didn't know where this data came from myself until I found you actually found this great study from 1971. And so I I'll I'll happily review it. So thanks for bringing this to my attention.

So this is a study by Dr. Petz in 1971. It was uh basically a quote in his words, quote our own review of the published literature and case reports available to us, unquote. And basically what he did was he summed the results of these uh these studies into kind of like uh pseudo-meta-analysis And what he found was that uh in patients who were not allergic to penicillin, uh they had a baseline rate of allergic reactions to cephalosporins approximately 1.7 % of the time, meaning 1.7% of patients had an allergy to cephalosporins just from the general population. However, in patients who had a documented penicillin allergy, the rate of allergic reactions to cephalosporins increased to 8.2%.

And again, so that's kind of where the number might be coming from, this meta-analysis type of study from 1971 So it seems like a lot of penicillin allergic patients are getting cephalosporin safely. What do you think could account for this discrepancy Excellent question. There are a number of hypotheses, um some more substantiated with evidence than others Uh it was actually very nicely reviewed in a recent uh uh publication Ansium Analytica by Dr. Vorobajek and colleagues And they basically kind of outline four hypotheses, and I'll emphasize these are hypotheses and I'll go over evidence to support one of them. So the first reason that's speculated why administration of stepzolin to patients with a penicillin allergy is safe is because patients with a pen reporting a penicillin allergy are in fact not reporting a true allergic reaction, an immune-mediated response to the penicillin molecule.

They may be reporting something like a GI upset or RASH but nothing that is actually related to a type 1 hypersensitivity reaction from an IgE antibody binding mast cells, and that would be the thing that would be at greater risk So that's the first reason. The second reason is even in patients who have reported a penicillin allergy, then they have known documented antibodies against penicillin, many of those antibodies will in fact dissipate over time. Such that as they grow into adulthood, if they had a prior appenticillin allergy as a child, many of those antibodies are in fact vanished. And I'll go over one study in a little bit that'll kind of talk about how that is.

The third reason that uh having a penicillin allergy may not lead to uh an adverse result from receiving cephalin is that the uh R chains and R groups are in fact significantly different So with penicillin, it has its set of R chains, and then it only actually relates the penicillin R chains only really cross-react with maybe like five or six different commercially available cephalosporins. Which leaves dozens and dozens of cephalosporins which do not have an arc chain that is related to penicillin And in fact, cephozoin specifically, which has two R chains like all the cefalsporins, does not share any R-chain similarity. at all with any of the other cephalosporines nor with penicillin, implying that antibodies to a uh uh the uh R chain moiety of penicillin would not recognize the R chain moiety of cephalin.

Now to emphasize this particular hypothesis is almost completely theoretical. While it is well known that the R chains for these two molecules are quite different They do share a beta-lactam backbone, so theoretically, if someone were to have uh IgE antibodies against a common beta-lactam backbone, That may trigger a cross-reactive reaction. But the speculation is because the antibodies may be directed against the R chains, that these patients would not have an cross-reaction. Matt, I was gonna have you verbally describe these R chains, but I think we're better off if I just put them on the website. That sounds like a fantastic idea.

I'm not sure I could describe them off the top of my head either. And I think you got one more hypothesis, right? So in many of the studies that Dr. Petz mentioned in his review from the 1970s that cited a cross-reactivity between penicillin and cephalosporines The preparation of those cephalosporins actually came from the native fungus which produced the cephalosporins. And as part of that preparation, the fungus would also produce some penicillin. So the purified cefalosporin from the fungus would have penicillin contamination.

Thus, if you were to administer that preparation to a patient Who had a known penicillin allergy, that patient would have an allergic reaction because you administered them a contaminated preparation with penicillin to which they have an allergy. Very straightforward. In today's era, cephalosporins are created not from fungus balls, but instead from synthetic chemistry. So this synthetic chemistry does not have this uh problem with uh penicillin contamination And thus, at least theoretically, it's thought that the many patients who had this quote cross-reactivity when were actually against a penicillin contaminant would not have a reaction to these other preparations Do you have any evidence to support this theory? Yeah, so good question.

So basically this is like the lowest level of evidence, which is the equivalent of email in 1970. So what happened was in the initial studies in the 70s, uh others raised the possibility that the uh allergic reaction of cephalosporins could be from penicillin contamination. So what the authors did from those studies was to uh write to the drug manufacturers who are manufacturing the cephalosporin preparation and ask them. How contaminated are your preparations of cephalus porn? And the drug manufacturers promptly and kindly responded that they were completely within the FDA limits. which said that you were allowed to have up to 0.05 units of penicillin per gram of cephalophin.

So not zero. So the researchers then did follow-up studies basically administering what would be what was thought to be the contaminated amount of penicillin in the cephalosporin preparations directly to patients in this very low doses. They did a dose escalation curve. And so what the researchers found was they believed that to get the type of allergic reaction they were seeing, you would need to administer on the order of 1,000-fold more penicillin Then what would be present in the preparation of cephalosporins that the manufacturers were claiming? The downside is uh it's unclear the degree of quality control that was going into this and how much oversight the manufacturer had during the manufacturing of the Cephalsporm preparations.

So we just don't know. But like I mentioned, this is low-level evidence. So the study basically concluded that at the level of contamination that these manufacturers were reporting it's unlikely that a clinically significant reaction would occur. Precisely. Taken at its face value, this would suggest that the difference It's between the lower rates of cross-reactive allergic reactions in today's population compared to prior population is unlikely to be accounted for by the uh contamination of penicillin with prior preparations of cephalosporins.

There's likely to be the alternative explanations. I think that's a really great review of some of the hypotheses that have been put forth by people about why we have some of the observations that we do nowadays in the data. Yeah, exactly. I'll just go into a little bit of depth on one of the studies. So um one of the major uh understandings about why patients with a penicillin allergy, particularly those reported as child uh don't have uh allergy to cephalosporins later in life is that the antibodies against penicillin dissipate over time So this is a fairly old study from the 70s by Think et al. And what they did was they basically just had a large group of patients, approximately 180 people, who had a known penicillin allergy And then they took them and they just skin tested them against penicillin.

And so what they found was that in this cohort of about 120 patients, um about 56% of patients who had had a reaction to penicillin in the last three months had a skin test that was positive to penicillin. So they had so that was 56%. They then looked at patients who reported a reaction that was further out, either a one year out, three years out, five years, ten years, twenty years. And what they found is as time progressed the proportion of patients who had a positive skin test to a penicillin challenge became lower and lower and lower uh to the point where it was approximately 12% at 10 years and approximately 0% at 20 years, suggesting that there is some time element that's important to developing allergic reaction And at least physiologically this makes sense is what we know about immunology.

That uh as time passes, if you are no longer exposed to antigens those uh B cells that produce the antibodies will dissipate and similarly the antibodies themselves will also dissipate. And this is why for instance when you have immunizations you may need intermittent booster shots. Like did you get your tetanus shot for every 10 years? It's because those B cells and that antigen is no longer present, the B cell dies, the antibody goes away. The thought is similar with the penicillin molecule as well.

If you're not exposed to it, that antibody will also go away. You might not believe this, but I have a documented penicillin analogy and it's also from when I was a kid. I believe it, and I don't believe it's real. Well, if what you're telling me is true about these IGE antibodies, I can finally eat that fungus ball penicillin thing that I've been looking forward to. It's a good thing.

Uh I actually have a loaf of bread that's been up there since uh about September. I'll I'll make sure to come back and bring it for ya. And I think the last claim that actually spontaneously came up as as we were talking about some of this data is that there's a concept of standard of care. So Some anesthesiologists as the rationale for why they won't give penicillin allergic patients cephalosporins is that say an anaphylactic reaction happens. would they be able to back that up in court? So I think we can identify that as the third and final claim being that in penicillin allergic patients, if you were to give a cephalosporin and anaphylaxis resulted, you could be found liable.

Yes, this is a great paper. I'm actually glad that you uh you found this. I didn't even realize this paper existed. So this is published uh in the uh Annals of Allergy, asthma, and immunology in 2018 And it's by a group that looked at a closed claims database of claims made against position as not practice lawsuits. And so what they were looking at was in patients who reported a penicillin allergy and then had a negative outcome as a result of receiving an antibiotic Um, what would happen if that patient or their family sued the physician who ordered that antibiotic to be administered?

And what they looked at was that there were 27 cases in this closed claims database where a patient with a penicillin allergy received a beta-lactam and had a negative reaction. And what they found was uh in patients who had a penicillin allergy and received a penicillin-based antibiotic Fewer than half of the defendants were found liable in those cases. So seven patients had penicillin allergy, received a penicillin-based antibiotic. And only three of those cases resulted in judgments against the physician, so under 50%. And so suggesting that this isn't necessarily a slam dunk reason to get sued Furthermore, in patients who had a penicillin allergy but instead received a cephalosporin or a carpenum antibiotic, so a related but not identical drug to the reported allergy Um there were no judgments found against physicians in court.

And in fact there was a a single case that was excluded but that was settled out of court, but of the remaining cases, none of the judgments were found against the the physician. And what I found especially interesting in this study was that the judges specifically cited, and I quote, a lack of scientific evidence demonstrating cephalosporins or carbapenins are contraindicated for patients with a penicillin allergy. Now again, this is a very small study, so there there are only a s a toll of seven cases. Um and uh so it's hard to extrapolate this necessarily to a very large generalizable population. Nonetheless, at least in the small meta-analysis, suggests that the law would be on the side of the physician who administers cephalosporin to a penicillin allergic patient.

I think that's super interesting, and it's not often that we get papers talking about professional liability, and we're really lucky that that one directly addresses the third claim that we just talked about. To conclude our show today, I'd like to just review some of the claims that we discussed today. The first being that in penicillin allergic patients, an alternative non-beta-lactam antibiotic may be just as effective. The second claim is that in penicillin allergic patients, administering a beta-lactam antibiotic like cefazolin may be unsafe due to the potential for cross-reactivity. And the third claim being that in penicillin allergic patients, administering a cephalosporin may be against standard of practice and subject you to professional liability.

To our listeners out there, what's the level of your belief in these claims? And again, what's the level of evidence supporting these claims? Matt, thanks for joining us on the show. I thought this was an excellent review of a very pertinent topic for anesthesiologists. Thanks for coming and I hope to have you back on the show again.

Pleasure's mine, and I hope to be back another time soon Thanks again for joining us this week. We had a lot of fun making this episode and we can be reached at depth of anesthesia at gmail. com and depth of anesthesia. com. You can also get in touch with us on Twitter We share interesting articles through that account and we also conduct some pretty interesting polls. Most recently we actually conducted a poll asking people whether or not they check mask ventilation before administering a paralytic and the results were yes at twenty eight percent, no at fifty four percent, and depends at eighteen percent, and that was out of a hundred thirteen votes. If you enjoyed the show this week, please leave us a five star review on iTunes and share this podcast with anybody who's involved in anesthesia and interested in learning more about the evidence behind why we do the things that we do.