Penicillin allergy in the hospital

David McCullagh, Derek Chu, Roman Jaeschke


McCullagh DJ, Chu DK. Penicillin allergy. CMAJ. 2019 Feb 25;191(8):E231. doi: 10.1503/cmaj.181117. PubMed PMID: 30803953; PubMed Central PMCID: PMC6389453.

Roman Jaeschke, MD, MSc: Good afternoon, welcome to another edition of McMaster Perspective. I have the pleasure of seeing fellows specializing in allergies and infectious diseases, Doctor Derek Chu and Doctor David McCullagh. I would like to talk first about penicillin allergy in the acute setting, in hospitals. That is the dilemma I face sometimes, when in the middle of the night comes a patient who would benefit from either a penicillin, beta-lactam antibiotic, or carbapenem, but somebody mentions this patient has a history of allergy to penicillin. What would you advise in this situation?

David McCullagh, MD: I think the first thing we should say is that under no circumstances should you delay any urgent, potentially life-saving or mortality-improving treatment. So it would be important to give the patient appropriate antibiotics before you delve into any form of significant history taking.

When you are taking the history—and that is the first thing you should do; you can take it from the patient or a relative—you are looking to put the patient into one of 3 different (roughly) risk categories. The first category is a patient with a similar risk of penicillin allergy as the general public. These would be people who have just an isolated side effect to a penicillin, such as a gastrointestinal (GI) side effect, like nausea, vomiting, diarrhea. If you can confirm that is the only reaction, then you can be reassured that this person has a similar risk of a significant allergy as the general public.

The other [factor] would be a family history. Some people will remark that their family history makes them at higher risk of penicillin allergy. In fact, that is not the case and they would be at similar risk as the general public, too.

The second risk category would be people in whom a penicillin or beta-lactam would be significantly or absolutely contraindicated. These would be people who have a severe delayed hypersensitivity reaction, such as a drug reaction with eosinophilia and systemic symptoms, toxic epidermal necrolysis or Stevens-Johnson syndrome, people who have had acute interstitial nephritis, serum sickness, or joint involvement with their allergy in the past. Those would be a big red flag to avoid penicillins or beta-lactams.

The other people that would be in a high-risk category would be people who have had a recent—in the last few months to a few years—significant anaphylactic [reaction], so a serious immediate hypersensitivity reaction. These are the people whom I would put in a high-risk category for a significant allergic event.

The more interesting category for us and the more difficult to determine is an intermediate category. These would be people who have had an immediate hypersensitivity reaction, but a distant one. We know that at 5 years people who have had a true IgE-mediated allergy will only have a 1 in 2 chance of continuing to have this allergy. By 10 years, it will only be an 80% chance of having this allergy.

Roman Jaeschke: You mean a 20% chance. Eighty percent less. After 10 years, 80% will lose the allergy.

David McCullagh: That is exactly right. These people would be an intermediate group. And then a big group that you get in hospital medicine is people who say “I cannot remember,” or for some reason you cannot identify what that allergy was. I would put those people in an intermediate group, too. The decision-making about what antibiotic to give and how to further evaluate this patient becomes a little bit more difficult.

Roman Jaeschke: Maybe we will use Dr Chu’s experience now. Is there any place in this intermediate category when we are not sure to use a beta-lactam? First, what is the risk that people with reported allergy to penicillin have beta-lactam allergy? And second, how can we push ourselves to start those antibiotics, even if we are not 100% sure?

Derek Chu, MD, PhD: I think there is 2 key points that your question raises. Number one is that previously penicillin allergy would be thought to encompass all beta-lactams, but there is more and more information coming out that is saying penicillin allergy is its own entity, cephalosporin allergy is its own entity, carbapenem allergies are its own entity, and that in general cross-reactivity between each of the categories is actually quite low. To use an example from penicillins, if one is allergic to ampicillin, they would have a 2% or less chance of cross-reacting with a cephalosporin, such as cefazolin. Same thing for carbapenems: it is about a 1% chance of cross-reactivity. So in general it is thought to be quite safe to use cephalosporins or carbapenems.

The second question is how to do it safely. There are 2 main ways of administering penicillins to a potentially allergic patient in an acute situation, where you do not have time. The first consideration is clinical stability. The patient must have sufficient monitoring as well as blood pressure support to ensure that you will not drop their blood pressure from eliciting a possible iatrogenic allergic reaction. The first of the 2 ways is called test dose, where one administers 1/100 or 1/10 of the full dose to the patient, observing them after each of those doses before administering the full dose. The second is desensitization, which is a roughly 7- to 12-step procedure where you provide incrementally increasing doses, so that the patient temporarily loses their allergy.

Roman Jaeschke: How do you perform desensitization? Within a period of time? At intervals?

Derek Chu: Exactly. It is typically in 20- to 30-minute intervals. You can monitor the patient for another 20 to 30 minutes thereafter, or longer if there are any signs of instability or reaction.

Roman Jaeschke: And you start from roughly 1/100 of the dose and then escalate it?

Derek Chu: There are different protocols. For a full desensitization procedure, say, one that is 12 steps, you may start as low as 1/100 or even 1/1000 and then increase at logarithmic or semilogarithmic folds. Often this will be guided by an allergist.

The test dose is the one that would be most readily accessible. For example, if someone reports a penicillin allergy, they were not recently hospitalized for it, and they did not have an absolute contraindication, such as Stevens-Johnsons syndrome being their reaction—that is, being hospitalized to the intensive care or burn unit because of a drug reaction—then what you could do is you could start with a separate beta-lactam of a different class, such as a cephalosporin or even a carbapenem, at 1/100 of the dose. For instance, instead of using ceftriaxone 1 g, you start at 10 mg, the patient receives it over 20 to 30 minutes, you watch for 20 to 30 minutes, and if there is no reaction, you escalate to 1/10 of the dose, so that you are at 100 mg, and then the full dose.

Roman Jaeschke: Maybe we will stop here with the use in the acute setting and I will ask you for a separate session on more chronic situations. For the time being, thank you very much. We will talk again soon.

See also

We would love to hear from you

  • Do you have any comments?
  • Have you found a mistake?
  • Would you like to suggest a feature?

We use cookies to ensure you get the best browsing experience on our website. Refer to our Cookies Information and Privacy Policy for more details.