A new tool to help identify true penicillin allergies

EBM Focus - Volume 15, Issue 13

Reference: JAMA Intern Med. 2020 Mar 16

Patients often self-report penicillin allergies, and when inaccurate, this can interfere with use of appropriate antibiotics. Although formal antibiotic allergy testing can be incorporated into good antibiotic stewardship, it is labor intensive, requires expert interpretation, and is of limited utility given the scarcity of allergists who incorporate drug allergy testing into their practice. In contrast, a validated, straightforward clinical decision rule could be used at the point of care by both specialists and non-specialists.

A multicenter prospective diagnostic cohort study was recently conducted to develop and validate such a clinical decision rule. This study completed derivation and validation with a cohort of 622 patients (median age 60 years, 59% female) from two tertiary care sites in Melbourne, Australia. Researchers then applied this rule retrospectively to 945 patients from Sydney, Perth, and Nashville (median age 55, 70.1% female) who had reported penicillin allergies and subsequently underwent testing. Patients reporting a penicillin allergy had skin prick, intradermal, patch, and/or oral challenge testing either directly or after skin testing, and patients with a negative skin prick or intradermal result without oral challenge were excluded.

In the derivation and validation cohort, 9.3% of patients reporting a penicillin allergy had a positive allergy test. The four characteristics associated with a positive test and used to develop the clinical decision rule included penicillin allergy reported by the patient (not scored), five or fewer years since last reaction (2 points), either anaphylaxis/angioedema or severe cutaneous adverse reaction (2 points), and treatment required for allergy episode (1 point) (or PEN-FAST). Scores were stratified as follows: 0 points - very low (0.6%) risk, 1-2 points - low (5%) risk, 3 points - moderate (20%) risk, and 4-5 points, high (50%) risk of positive penicillin allergy test, with a cut-off of 3 points chosen to dichotomize lower vs. higher risk. This cut-off had sensitivity 70.7%, specificity 78.5%, positive predictive value (PPV) 25.3%, and negative predictive value (NPV) 96.3% in the internal derivation cohort. In the external validation cohorts, positive penicillin allergy tests ranged from 3.6% to 33.8% depending on site. This contributed to a wide range for the PPV (6.4% to 70.4%) although the NPV (84.9% to 98.4%) was not affected as strongly.

Overall, PEN-FAST shows good diagnostic performance. There was considerable variation in the performance between the external validation cohorts, but this is likely due in part to the varying prevalence of positive test results. The authors comment that this rule should not be generalized to pediatric or ethnically diverse populations. Other acknowledged limitations included the exclusion of beta-lactam allergies and IV penicillins, a disproportionate percentage of inpatient testing, a relatively small sample of patients with severe cutaneous adverse reactions, and a lack of documentation of treatment of acute allergy episodes in the external validation cohorts. Finally, the NPV suggests that up to 15% of patients who are negative by PEN-FAST will still have a penicillin allergy, so proceeding cautiously at first is still warranted. Nevertheless, PEN-FAST represents a speedy, straightforward clinical decision rule that can be used by both generalists and specialists alike to better serve patients reporting penicillin allergies.

For more information, see the topic Penicillin Allergy in DynaMed.

DynaMed EBM Focus Editorial Team

This EBM Focus was written by Terri Levine, PhD, Senior Medical Writer in Obstetrics and Gynecology at DynaMed. Edited by Alan Ehrlich, MD, Executive Editor at DynaMed and Associate Professor in Family Medicine at the University of Massachusetts Medical School, Dan Randall, MD, Deputy Editor for Internal Medicine at DynaMed, and Katharine DeGeorge, MD, MS, Associate Professor of Family Medicine at the University of Virginia and Clinical Editor at DynaMed.