An estimated 10% of the U.S. population report having a penicillin allergy. Most were given the label in the course of a childhood health care encounter —75% by age 3.
According to the U.S. Centers for Disease Control and Prevention, fewer than 1% of the population is truly allergic to penicillin. The rest were never allergic to begin with or have outgrown their allergy — an estimated 80% of people with penicillin allergy lose their sensitivity to the drug within 10 years.
Outgrown or misplaced penicillin allergy labels contribute to overuse of alternative, broad-based antibiotics and the development of drug-resistant bacteria. Inability to give penicillin to patients in the hospital results in a higher incidence of drug-resistant infections, longer hospital stays and higher costs.
“Penicillin allergy labels that don’t hold water are a significant public health issue,” said infectious diseases specialist Elizabeth Phillips, MD, the John A. Oates Chair in Clinical Research and professor of Medicine at Vanderbilt University Medical Center.
Phillips and colleagues are making headway on this issue.
- Using a penicillin allergy risk profile developed at VUMC, allergy and immunology specialist Cosby Stone, MD, MPH, instructor in Medicine, is leading an effort to remove labels from inpatients, using a so-called oral challenge in the form of a mild, 250 mg dose of amoxicillin (a type of penicillin). As of April 24, 100 patients admitted to the medical intensive care unit have had labels removed. A recent report on this effort appears in the American Journal of Respiratory and Critical Care Medicine.
- In a new survey led by Allergy and Immunology Fellow David Coleman, MD, patients who reported penicillin allergy were more likely to report use of broad spectrum and second-line antibiotics; some 80% of patients who reported having the allergy believed it to be permanent; and although most with the allergy were willing to undergo testing, less than 5% had been offered a referral for this testing. The survey is reported in the Journal of Allergy & Immunology: In Practice.
Phillips, the senior author for both of those reports, has now published a study with researchers in Australia, helping to put risk-profiling for penicillin allergy on an international footing. Their report appears in JAMA Internal Medicine.
“Using a small set of criteria, we identify a decision strategy that allows allergists and non-allergists alike to risk-stratify these patients and direct appropriate delabeling and prescribing,” Phillips said.
“For inpatients with the label who are found to in fact be at low risk for the allergy — those with scores below 3, per our decision rule — the hospital team can safely proceed directly to an oral challenge, circumventing the skin testing employed by allergists in the outpatient setting.”
In the study’s primary cohort, 622 patients in Melbourne who reported penicillin allergy underwent testing and 9.3% tested positive for the allergy. Medical histories taken from this cohort were used to construct a predictive model for a positive penicillin allergy test result. External validation of the model was performed in retrospective penicillin allergy–tested cohorts consisting of 945 patients from Sydney, Perth and Nashville.
The resulting decision rule involves scoring of the following patient features: a penicillin allergy event occurring five or fewer years ago, 2 points; a history that includes a severe allergic reaction to penicillin, including anaphylaxis, angioedema or severe cutaneous adverse reaction, 2 points; a history that includes an allergy episode requiring treatment, 1 point.
Some 74% of patients in the primary cohort wound up with scores below 3.
Per the decision rule, a score of zero indicates a below 1% risk of a positive penicillin allergy test; a score of 1 or 2 indicates a risk of 5%; a score of 3 indicates a risk of 20%; a score of 4 or 5 indicates a risk of 50%.
Other VUMC researchers on the study include Stone and Roger Yu, a research analyst with the Division of Infectious Diseases. The study was supported in part by grants from the National Institutes of Health (GM115305, AI139021).
