A woman with a history of anaphylaxis to penicillin is admitted with cholecystitis. Which of the following is the best regimen for her?
- Levofloxacin and metronidazole
- Vancomycin, Aztreonam, and metronidazole
The first two options aren’t perfect:
- Levofloxacin + metronidazole: Fluoroquinolones aren’t great in the ICU for several reasons, including increasing resistance against gram-negatives and adverse events.
- Vancomycin + aztreonam + metronidazole: For most gram-positives, vancomycin is a sub-optimal antibiotic. It is a large molecule with impaired penetration of some spaces. It is nephrotoxic, with a limited therapeutic index (often leading to either inadequate or toxic levels). Thus, for a patient who doesn’t need MRSA coverage, vancomycin as the sole agent to cover gram-positives isn’t ideal.
Meropenem is arguably the best option, but there may be concern about the risk of another anaphylactic reaction. Is it safe?
Meropenem safety: Examining the evidence
Meropenem doesn’t cause anaphylaxis
There don’t seem to be any clear reports of meropenem causing anaphylaxis. Pubmed search reveals no case reports. A meta-analysis of studies including >6,000 patients exposed to meropenem found no case of anaphylaxis (Linden 2007). The closest I could find was passing mention of one patient who supposedly developed anaphylaxis three days after starting meropenem (1).
Retrospective studies describe 838 patients with penicillin allergy who were treated with meropenem (Kula 2014). According to a consensus guideline:
Retrospective studies of hospitalized patients with a history of penicillin allergy…showed that approximately 10% developed possibly allergic reactions during treatment with carbapenems, and none of these reactions was life-threatening (Multi-society guideline, 2010)(2).
We’re afraid to use meropenem due to a fear of anaphylaxis. Currently there is little evidence that this even exists. Of course, anything is possible. However, there is stronger evidence that people are killed by entanglement in bedsheets than by anaphylaxis from meropenem (Kibayashi 2011). Therefore, it would be more logical to remove bedsheets from the hospital than to avoid meropenem due to fear of anaphylaxis.
Meropenem probably isn’t cross-allergic with penicillin
It used to be feared that if patients were allergic to the beta-lactam core structure, they would be allergic to every beta-lactam antibiotic. Fortunately, this isn’t the case. Patients aren’t allergic to the beta-lactam core, but rather to the different side-chains that each antibiotic has. Some antibiotics have similar side-chains, making them are cross-allergic (e.g. amoxicillin and cephalexin). Other antibiotics have unique side-chains which don’t cause cross-allergic reactions.Most studies assume that meropenem is cross-allergic with penicillin. Probably not. Wall 2014 performed a large study to test this concept. 914 patients were retrospectively identified who received a carbapenem. Although underpowered, there was trend towards fewer allergic reactions in the patients with a history of penicillin allergy (1/324 vs. 4/634).
Studies based on skin testing may overestimate risk
It is widely quoted that if patients are allergic to penicillin, there is a ~1% risk that they will be allergic to meropenem. This number is derived from a study of patients with a positive skin test for penicillin allergy (Romano 2007). Among 103 such patients, one had a positive skin test for meropenem. Patients with a negative skin test were challenged with IV meropenem and tolerated it well, whereas the patient with a positive skin test wasn’t challenged.
This study seems to imply that the patient with a positive skin test would have had a horrible reaction to meropenem. However, the specificity of skin testing to meropenem is unknown. It’s possible that this widely quoted “1%” figure is merely a measurement of the false-positive rate of unvalidated skin testing with meropenem (3).
Gaeta 2015 performed a larger study with similar design. 212 patients with a history of penicillin allergy and a positive skin test for penicillin were included. Patients had allergies to various beta-lactams (penicillin, amoxacillin, ampicillin, piperacillin, and occasionally cephalosporins). No patient had a skin test reaction to meropenem. All patients were challenged with IV meropenem and tolerated it well (4). Notably, 77% of these patients had a documented history of anaphylaxis to penicillin.
Prospective validation of meropenem use among patients with penicillin anaphylaxis
Cunha 2008 performed a prospective study describing the use of meropenem among 110 patients with penicillin allergy (51 of whom had a history of anaphylaxis). Patients were treated using meropenem, without prior skin testing or graded challenge. There were no allergic reactions. The low rate of allergic reactions in this study may reflect the exclusion of patients on multiple antibiotics (when using multiple antibiotics, it’s difficult to attribute an allergic reaction to any specific drug).
The weakness of this study is that it is underpowered to exclude rare adverse reactions. However, these investigators have continued to use meropenem aggressively, and after nearly a decade of additional experience they still haven’t had any problems with it:
We have continued to treat hundreds of patients with meropenem with a history of anaphylactic and non-anaphylactic reactions to penicillins, even after prolonged therapy for days or weeks, and have never had any allergic reactions. –Cunha BA 2015
What do the guideline say?
American Allergy societies recommend that before using any cephalosporin or carbapenem, the patient should have a skin test and/or graded challenge with the drug. This isn’t logistically feasible for a septic patient, making it a bit of a copout (figure below). The only drug which these guidelines recommend without a skin test or graded challenge is aztreonam.
Approaches used by the Infectious Disease Society of America are inconsistent. The 2016 VAP and 2007 CAP guidelines both use aztreonam for patients with penicillin allergy. However, the 2015 endocarditis guidelines recommend cefazolin for patients with a non-anaphylactic penicillin allergy.
Below is my general approach to patients allergic to penicillins (e.g., penicillin, ampicillin, piperacillin). The rate of cross-reaction between penicillins and cephalosporins with dissimilar side-chains is extremely low, so most of these patients can be treated with a cephalosporin (Campagna 2012). For patients with a history of anaphylaxis, aztroenam or meropenem may be slightly safer (5).
Please note that cefazolin isn’t the only first generation cephalosporin that could be used. Depending on exactly which drug the patient reacted to, various other cephalosporins might be considered (this depends on the relationship of side-chains on various beta-lactams)(6).
- Meropenem doesn’t cause anaphylaxis.
- Meropenem probably isn’t cross-allergic with penicillin, due to structural differences in its side chains and beta-lactam core.
- It is widely quoted that patients with penicillin allergy have a ~1% risk of reaction to meropenem. This information is based on non-validated skin testing to meropenem, which is probably incorrect.
- Prospective studies have proven the safety of giving meropenem to patients with a history of anaphylaxis to penicillin.
- Cunha BA 2015 No need for an initial test dose of meropenem or ertapenem in patients reporting anaphylactic reactions to penicillins. J Chemotherapy 27(5) 317.
- Dr. Cunha is an international leader in infectious diseases with >600 publications in Pubmed. This is a good citation if you need something to back up your use of meropenem in a patient with PCN anaphylaxis.
- Cephalosporins in PCN allergy
- Mythbuster: 10% cephalosporin-penicillin cross-reactivity (Bryan Hayes, ALIEM)
- Cephalosporins and penicillin cross-reactivity (Rebel EM)
- Cephalosporins can be used in PCN allergy (EM Literature of Note)
- Low utility of fluoroquinolones in critical care (PulmCrit)
- See Table 4 in Prescott 2004. One 27-year-old patient with a history of rash to penicillin in this study supposedly developed anaphylaxis three days after challenge with meropenem. This time delay is inconsistent with the natural history of anaphylaxis. Many patients in this study were receiving multiple antibiotics. Without any further details regarding this case and exactly which medications this patient was exposed to, it’s difficult to conclude that this was anaphylaxis from meropenem.
- More recent evidence suggests that this 10% figure is too high (Kula 2014). It's probably equal to zero. Regardless, all of these studies agree that these are minor reactions.
- The positive predictive value of a penicillin skin test is 50%, so it seems that a positive skin test to penicillin isn’t perfect either (Gonzalez-Estrada 2015). The sensitivity and specificity of a meropenem skin test are unknown, but it seems reasonable to assume that the test isn’t 100% specific.
- Except for one patient who declined receiving meropenem.
- My guess is that it would be fine to use a third/fourth generation cephalosporin in a patient with a history of anaphylaxis to penicillin. However, currently there doesn’t seem to be enough evidence to solidly back this up. Hopefully such evidence will emerge in the near-future. Of course, every patient must be approached individually with attention to the allergy history (e.g., how severe the reaction was, how long ago it was, etc.) and the risks/benefits of different antibiotic strategies.
- For inpatients who are getting IV antibiotics, there doesn’t seem to be an advantage of using a different first-generation cephalosporin instead of cefazolin. Thus, this algorithm is sufficiently complicated to serve my needs. Depending on your scope of practice and the antibiotics available to you, you might need a more complex algorithm.