D1.07 - Risk Stratification Tools in Suspected Beta-Lactam Allergy: A RealLife Delabeling Study

The incorrect labeling of β-lactam allergy is a major clinical and public health issue. Although up to 25% of hospitalized patients report penicillin allergy, fewer than 5% are confirmed by standardized diagnostic testing. This discrepancy results in inappropriate antibiotic use, increased healthcare costs, and contributes to antimicrobial resistance. This study evaluated the diagnostic performance of three validated risk stratification tools - PEN-FAST, the 1:1:1 criterion for urticaria, and the EAACI/ENDA algorithm - in a real-world cohort of patients with suspected βlactam hypersensitivity, and assessed outcomes after delabeling.

We conducted a retrospective observational study including 140 patients evaluated between 2021 and 2025 at the Center for Personalized Medicine in Asthma and Allergology, Humanitas San Pio X Hospital (Milan, Italy). Forty-seven patients with a complete diagnostic work-up were included. PEN-FAST, 1:1:1 criterion, and EAACI/ENDA classification were retrospectively applied and compared with standard diagnostic outcomes (specific IgE, skin tests, and drug provocation tests). Diagnostic performance was assessed by sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). A prospective followup evaluated re-exposure and confidence in β-lactam reuse among delabeled patients.

In this real-world cohort, the EAACI/ENDA algorithm provided the best balance of sensitivity and safety, while the 1:1:1 criterion reliably identified low-risk profiles. PEN-FAST showed limited discriminative capacity. No model performed uniformly, underscoring the need for context-specific application. Integrating structured risk stratification into antimicrobial stewardship programs may optimize β-lactam use and improve patient safety.

In this real-world cohort, the EAACI/ENDA algorithm provided the best balance of sensitivity and safety, while the 1:1:1 criterion reliably identified low-risk profiles. PEN-FAST showed limited discriminative capacity. No model performed uniformly, underscoring the need for context-specific application. Integrating structured risk stratification into antimicrobial stewardship programs may optimize β-lactam use and improve patient safety.