Infectious Disease Compendium

Antibiotic Allergy


Most patients who think they are allergic to penicillin are not (from):

"Only 10% to 20% of patients reporting a history of penicillin allergy are truly allergic when assessed by skin testing. Taking a detailed history of a patient's reaction to penicillin may allow clinicians to exclude true penicillin allergy, allowing these patients to receive penicillin. Patients with a concerning history of type I penicillin allergy who have a compelling need for a drug containing penicillin should undergo skin testing. Virtually all patients with a negative skin test result can take penicillin without serious sequelae."

And most of the history's taken are inadequate. The patient says they are allergic to a given antibiotic. Ask what the allergy was. Nausea and vomiting is not an allergy but an adverse reaction .What you worry about are Type I Immediate Hypersensitivity Reactions. It is medicated by IgE and the one that kills from anaphylaxis: urticarial rash (hives), itching, flushing, angioedema, wheezing, hypotension, progressing to death.

And in the old days it was contamination that led to the allergic reactions as well as penicillin break down products. I cannot find data to suggest if Type 1 antibiotic allergies fade with time. How often does penicillin allergy cross react with ampicillin? Reasonably often, but not always:

"Thirty subjects with a history of allergic reactions to penicillins were studied. In vivo and in vitro specific IgE antibodies were determined to different penicillin determinants. Fifteen subjects developed anaphylactic responses and the remainder urticaria and angioedema. The drug most frequently involved in the patients' allergic reactions was ampicillin (AMP). The benzylpenicilloyl (BPO) skin test was positive in 16 (53.3%) patients, whereas 23 (76.6%) patients were positive to minor determinant mixture (MDM), benzylpenicillin (PG), AMP, or amoxicillin. (AX)."

And as best I can tell, there are (understandably) no clinical trials giving ampicillin to penicillin Type 1 allergic patients.

I also put the code cart in the room when I challenge them with a beta lactam. My only superstition: put the emergency equipment in the patients room and you will never have to use it.

BTW: in the era of home antibiotic infusion, be aware of the spouses allergy history: "We present a case of an immediate allergic reaction in a penicillin-sensitive spouse of a patient receiving parenteral mezlocillin sodium therapy. A seminal level of 42 μg/mL of mezlocillin was documented by bioassay."

Here is a nice bit of penicillin history: (PubMed). Read it instead of looking at your Facebook page next time you are on the toilet.

Not giving a beta-lactam because of some half-assed antibiotic allergy leads to suboptimal treatment (Pubmed) and patients have more adverse complications (Pubmed). Oncology patients who get alternatives on beta-lactams do worse (PubMed).

A documented peniclllin allergy is associated with increased risk of MRSA and C. difficile (PubMed).

And giving alternative antibiotics for surgical prophylaxis instead of a beta-lactam leads to increased in surgical wound infections (PubMed).

It is pretty clear: if you give a second line drug for a disease where a beta-lactam is preferred, the outcomes will be worse. So take a careful history and give that beta-lactam if the history supports you.

Some pearls:

Penicillin reactions from the 60's and 70's may have been (note the word may) due to impurities, not penicillin.

The maculopapular rash often seen with amoxicillin means nothing for a future allergic reaction and may be more sensitive for mononucleosis than a monospot.

If allergic to penicillin, there is maybe a 10% cross reaction to cephalosporins (1st > 2nd > 3rd), maybe 30% to imipenem, but 0% to meropenem (PubMed) and aztreonam.

"The cross-reactivity between penicillins and carbapenems for IgE-mediated reactions is very low, but caution is still advised. Cross-reactivity rates may be higher between cephalosporins and carbapenems (PubMed)."

Then there are they other hodgepodge of rashes and allergic reactions.

- hemolytic anemia from penicillins and cephalosporins.

- thrombocytopenia from beta-lactams, vancomycin, linezolid, and sulfonamides.

- serum sickness from amoxicillin and sulfonamides.

- vasculitis from penicillins, cephalosporins, and sulfonamides.

- acute interstitial nephritis from beta-lactams (especially nafcillin and oxacillin), vancomycin, linezolid, rifampin and sulfonamides. Always check for urine eosinophils in acute kidney injury on antibiotics, although neither sensitive nor specific.

- Stevens-Johnsons from sulfonamides. (AIDS increases risk), tetracyclines, and dapsone.

- drug rash with eosinophilia and systemic symptoms (DRESS), is a severe type IV hypersensitivity reaction with fever, rash, and multisystem failure from sulfonamides.

- drug fever from any drug; one hint is the fever never drops to normal. Most fevers are up and down (including drug) but if the fever curve is flat line (continuous) look for a drug as the reason.

Epidemiologic Risks

Being on a drug. Reactions can occur on the first dose or the 1000th dose, but they often manifest around day 10-14. And do not forget all the other drugs that can cause allergic reactions; it is not just the antibiotics.

Empiric Therapy

If there is a compelling reason to give an antibiotic to which the patient is allergic, consider desensitization. Antibiotic hypersensitivity reactions and approaches to desensitization is an excellent review from 2014 and covers a large number of drugs. I have only done it once, a pregnant woman with tertiary syphilis. It took most of my day to accomplish. Total pain.

There is increasing popularity in skin testing before giving beta-lactams (PubMed). One method (PubMed) is to do a prick test, then an intradermal injection with penicillin and controls, followed by po amoxicillin before the beta-lactam is given. The vast majority of the time the patient has no issues, so I wonder about the time spent (looks to be about six hours) and expense for such a low yield of positives vrs just giving a penicillin (which I tend to do) with the code cart right outside the door. Have yet to need it.



Curious Cases

Relevant links to my Medscape blog

Allergic? I do not think so.

Say Yes to the DRESS

Last Update: 07/03/18.