Sex- and-Gender-Based Analyses of Penicillin Allergy Labels, Outcomes, and Delabeling Interventions

Abstract

Background: Penicillin allergy labels (PALs) affect approximately 10% of the population, yet over 95% of labeled individuals tolerate penicillin following delabeling interventions. For patients who are considered low risk, oral challenge (OC) and direct delabeling (DD) are increasingly favored over penicillin skin testing (PST), although definitions of low risk vary. Sex-based differences in PAL prevalence and clinical characteristics remain poorly defined. Females are more frequently labeled with penicillin allergy and receive antibiotics more often than males, yet sex-based differences in PAL prevalence, delabeling outcomes, and antibiotic utilization practices remain poorly characterized. A clearer understanding of how sex influences PAL reporting, delabeling strategies, and antibiotic selection is needed to optimize allergy evaluation and antimicrobial stewardship.

Objectives: This thesis examined sex-based differences in PAL, delabeling strategies, and antibiotic utilization. A systematic review and meta-analysis evaluated effectiveness and safety of DD and OC in low-risk patients and assessed sex and gender differences in reporting and outcomes. A descriptive analysis characterized OC protocols, hypersensitivity reactions, and reported predictors of adverse events. A cross-sectional study assessed associations between sex and PAL in an ambulatory care hospital in Ontario, including differences in β-lactam allergy reporting, reaction severity, and comorbidities. Finally, a cohort study evaluated whether sex modifies the association between PAL and antibiotic prescribing, including antibiotic class selection and use of second-line agents when β-lactams are indicated.

Methods: This thesis is based on research utilizing multiple diverse study designs, including a systematic review and meta-analysis, a descriptive epidemiologic analysis, and cross-sectional and cohort studies. A PRISMA-guided quantitative systematic review was conducted through February 2024 to identify studies evaluating direct delabeling (DD) or oral challenge (OC) in low-risk patients compared with no intervention, penicillin skin testing, or alternative strategies. Study quality was independently assessed by two reviewers. Random-effects meta-analyses were performed, with subgroup analyses conducted in the presence of substantial heterogeneity (I² >75%), and qualitative data were synthesized narratively. Data from included OC studies were further analyzed descriptively to characterize challenge protocols, adverse reactions, and reported predictors of hypersensitivity, with reactions classified by timing and pooled adverse event rates estimated using random-effects models. A STROBE-guided cross-sectional study of adults attending an ambulatory urgent medicine clinic in Ontario (2015–2024) examined associations between sex and penicillin allergy labels (PALs) using multivariable logistic regression. A retrospective cohort study of the same population assessed whether sex modified the association between PAL status and antibiotic prescribing, including antibiotic class selection and β-lactam use when indicated, using multivariable logistic regression with interaction terms; Firth’s bias-reduced methods were applied when separation occurred.

Results: Across 28 studies included in the systematic review and meta-analysis (2 randomized trials, 26 quasi-experimental), sex was reported in 86%, with females comprising 61% of participants; however, only 18% disaggregated outcomes by sex and no studies reported gender variables. OC was similarly effective to PST in randomized trials (RR 1.04, 95% CI 0.95–1.13), while quasi-experimental studies reported high delabeling rates for OC (~90%). Direct delabeling was possible in approximately 27% of those included with a study arm for DD. A descriptive epidemiologic analysis of 26 OC studies demonstrated a pooled reaction or non-delabeling rate of 4% (95% CI 3–6%), most commonly mild cutaneous reactions, with higher baseline allergy burden and shorter intervals since index reaction being identified as potential risk factors. In a cross-sectional study of 29,645 ambulatory care patients in Ontario, 9.4% had a penicillin allergy label (PAL), most of which were low risk; females were significantly more likely to have a PAL than males, report multiple medication allergies, and demonstrate sex-specific associations between PAL and comorbidities. In multivariable analysis, female sex, atopic disease, older age, and number of other medication allergies were independently associated with PAL. In the cohort study, PAL was independently associated with increased antibiotic use, while female sex was associated with lower odds of antibiotic prescribing. Effect modification by sex was not detected between PAL and antibiotic use, class selection, or guideline-recommended therapy.

Conclusion: PALs remain common and are disproportionately reported among females, despite high tolerability of penicillin following evaluation. Low-risk delabeling strategies, particularly OC, are safe and effective, yet sex- and gender-disaggregated outcomes are infrequently reported limiting understanding of clinical implications. Although female sex was independently associated with PAL and distinct comorbidity patterns, sex did not modify the relationship between PALs and antibiotic utilization. These findings highlight persistent gaps in sex- and gender-sensitive evidence across the management of patients presenting with infections that have a PAL.