Use of cephalosporins in penicillin-allergic patients

Let's start with a patient case.  A 71 year old patient arrives to the emergency room complaining of shortness of breath and sputum production that has worsened from when it started about three days ago.  She has a past medical history of diabetes mellitus type 2, myasthenia gravis, and atrial fibrillation and is taking metformin, pyridostigmine, prednisone, warfarin, and sotalol.  Her BP = 132/88, HR = 78, RR = 24, and T = 37.5 C and CXR reveals a left lower lobe infiltrate.  The diagnosis of pneumonia is made and as you begin to type orders for your standard ceftriaxone/azithromycin combination, you note that the patient has a penicillin allergy.  What is the risk of continuing this antibiotic regimen given the patient's allergy?

A recent study retrospectively looked at nearly 4,000,000 unique individuals within the Kaiser Permanente Health Care Program.1  They recorded baseline demographic data (including allergy history) and the incidence of new cephalosporin allergy reports within 30 days of a course of a cephalosporin.  This means that data was collecting measuring the incidence of a new cephalosporin allergy in patients with no reported allergies, allergies to penicillin, allergies to cephalosporins, and allergies to other drugs.  Here were some of the findings:

• 65,915 patients with a documented penicillin allergy received 127,125 courses of cephalosporins
• 1439 had a new cephalosporin allergy (2.2%)
• 3 had an anaphylactic reaction (0.005%)
• 3,313 patients with a documented cephalosporin allergy received 6,404 courses of cephalosporins
• 45 had a new cephalosporin allergy (0.01%)
• Zero anaphylactic reactions
• There were new cephalosporin allergies in 0.8% patients with no history of allergies to any antibiotic
• Another way of looking at this is that for every 135 patients with a penicillin allergy treated with a cephalosporin, there would be one new cephalosporin allergy, compared to treating patients with no initial allergy.
• The most commonly prescribed oral cephalosporin in this study was cephalexin and the most common parenteral cephalosporin was ceftriaxone.

Results from this study suggest that the risk of an allergic reaction to a cephalosporin in a patient with a penicillin allergy is very low.  Though retrospective in nature, these results are consistent with estimates of cross-reactivity between these drug classes in the more recent decades.  The warnings of cross-allergy in the first- and second-generation cephalosporins reaching up to 10% are from older studies with significant methodologic limitations.  Several sources claim that cephalosporins pre-1980 were contaminated with penicillin, likely increasing the cross-reactivity, but I was unable to verify these claims.

Issues affecting assessing cross-reactivity

Here are a few points to consider when assessing the risk of cross-reactivity between penicillin and cephalosporins.  

• Actually an allergy?  Most patients who report historical allergies to penicillin do not have a reaction to a penicillin skin test or to oral amoxicillin challenges.  One study gave the penicillin skin test to 500 patients who reported allergic reactions to penicillin (excluded reactions like Stevens Johnson syndrome, toxic epidermal necrolysis, hemolytic anemia, hepatitis, or nephritis as these are non-IgE-mediated and have an onset beyond the observation period of the penicillin skin test).  Of 500 patients with reported allergies, only 4 had positive skin tests (0.8%).  The remaining 496 with a negative skin test were given oral amoxicillin and only 4 of those had allergic reactions (all developed hives).²
• Patient misunderstanding of 'allergy'  Part of the uncertainty is that patients likely mistake 'allergy' for any sort of adverse event, rather than what 'allergy' means to someone with medical training.  This is evident by the following exchange being all too common.  Patient: "It made me really sick when I took it."  Clinician:  "What do you mean that it made you 'really sick'?"  Patient: "I had a bunch of stomach pain and some diarrhea - I was told never to take it again."  Clinician: sigh
• Actually penicillin?  Another issue is that if a reaction did occur, 'penicillin' might not be the culprit antibiotic.  Ampicillin, amoxicillin, and cephalosporins, have been commonly used since the 1970s and 80s.  It is possible that a patient may have had a reaction to one of these antibiotics, not penicillin, and have been inappropriately labeled as 'penicillin allergic'.
• Mechanism for allergy?  Immune reactions to penicillins  occur due to antibody sensitization to either the beta-lactam core (common to all beta-lactams) or the R- group side chains (only certain cephalosporins share R- groups).  
• Amoxicillin shares an identical R- group with cefadroxil and cefprozil (so avoid these in an amoxicillin-allergic patient)
• Ampicillin share an identical R- group with cefaclor, cephalexin, and loracarbef (so avoid these in an ampicillin-allergic patient)
• The specific part of the structure that is recognized by the immune system can determine the pattern of cross-reactivity

Back to the patient case

First off, the gut reaction in this case may be to directly avoid the allergy issue by giving a fluoroquinolone.  Unfortunately, a fluoroquinolone is undesirable in this patient due to the risk of dysglycemias, arrhythmias, and the boxed warning against use in myasthenia gravis (can cause severe exacerbations requiring ventilator support).  The patient should be assessed as to the nature of the 'allergic' reaction, when it occurred, what the actual causative agent was (really penicillin?), and if they have tolerated any other beta-lactam antibiotics.  If the patient has tolerated other beta-lactams, it is likely that an allergic reaction (if it actually took place) was due to the R- group of the penicillin.  Since no cephalosporin shares any of these identical structures with penicillin, the risk of an allergic reaction to ceftriaxone is likely comparable to the risk of giving ceftriaxone to a non-penicillin allergic patient.

Take home points:

• Probe patients about the specific reaction, the specific drug that caused the reaction, and if the patient has tolerated other beta-lactam antibiotics
• If a reaction is witnessed/suspected, make sure to document in the medical record the specific drug (don't put 'penicillin' if they have a reaction to 'cephalexin')
• Some first- and second- generation cephalosporins share R- groups with amoxicillin and ampicillin and may be more likely to cross-react
• Avoid fluoroquinolone antibiotics in patients with myasthenic gravis (boxed warning)

References
1.  Macy E, Contreras R.  Adverse reactions associated with oral and parenteral use of cephalosporins: A retrospective population-based analysis.  J Allergy Clin Immunol 2015;135(3):745-52.
2.  Macy E, Ngor EW.  Safely diagnosing clinically significant penicillin allergy using only penicilloyl-poly-lysine, penicillin, and oral amoxicillin. J Allergy Clin Immunol Pract  2013;1(3):258-63.

photo by Science Museum London