Outcomes of Prolonged Provocation Testing in Penicillin Allergy Evaluation

Penicillin allergies are frequently reported, yet up to 90% of cases are not true allergies upon further evaluation [1]. Drug provocation testing (DPT) remains the gold standard for diagnosis. Prolonged DPT may enhance diagnostic sensitivity [2] and patient confidence in future penicillin usage [3], though concerns about unnecessary antibiotic exposure remain. The latest European Academy of Allergy & Clinical Immunology (EAACI) position paper calls for further evidence to clarify its role [4]. In this prospective study, we aimed to evaluate our centre's multidisciplinary penicillin allergy evaluation programme, focusing on the safety and utility of a 5-day prolonged DPT for patients with delayed onset or unknown reactions. From March 2019 to May 2023, adults referred to the Singapore General Hospital outpatient allergy centre for suspected penicillin allergy were recruited, excluding those with severe cutaneous adverse reactions. At the initial visit, detailed clinical histories and patient backgrounds were recorded. Reactions were classified as ‘Immediate’ (within 6 h of antibiotic consumption) or ‘Delayed’ (after 6 h). On challenge day, skin prick testing (SPT) and intra-dermal testing (IDT) were conducted, following European Network of Drug Allergy (ENDA) and EAACI guidelines [5, 6]. Negative skin tests were followed by graded in-clinic DPT, conducted with divided doses every 30 min until full therapeutic dose. Patients were then observed in clinic for an hour; if no objective reaction occurred, an immediate reaction was considered excluded. For patients with index reactions with delayed/unknown onset, and/or unknown clinical history, delayed IDT reading was performed, and if negative, a 5-day at-home prolonged DPT was then implemented. Patients were monitored through follow-up calls to ensure compliance and document reactions. Suspected positive reactors were encouraged to submit photographs or return for in-clinic evaluation. A total of 321 patients were included. The most commonly investigated drug was co-amoxiclav (87.2%), followed by amoxicillin (10.0%). The index reaction occurred within 6 h in 78 patients (24.3%), after 6 h in 81 (25.2%), while 162 (50.5%) were unable to recall their onset. The most common index reactions were rash (46.7%), angioedema (24.3%) and itch (15.0%). Of the 320 patients who completed evaluation (Figure 1a), 17 (5.3%) had positive skin tests. The remaining 303 underwent in-clinic DPT, with 33 (10.9%) having positive reactions. One patient had a positive delayed IDT. Among the 200 patients proceeding to prolonged DPT, 34 (17.0%) had positive reactions, mostly occurring after Day 3 (70.6%) (Figure 1b). The most frequent reactions included rash (23, 67.6%), itch (19, 55.9%) and/or angioedema (9, 26.5%). None exceeded the index reaction's severity; all were managed conservatively without corticosteroids or hospitalisation. Overall, 85 (26.6%) patients were confirmed to have true penicillin allergy, while 235 (73.4%) were safely de-labelled. Secondary analysis showed that patients with positive prolonged DPT or delayed IDT were less likely to report urticaria (11.3% vs. 30.0%, p = 0.043) as their index reaction. Additional information about study methods and findings is available in the following repository: https://doi.org/10.5281/zenodo.14690499. In our study, prolonged DPT identified 34 (17.0%) additional cases of true penicillin allergy, increasing diagnostic yield by 40.0% compared to single-day DPT alone. These findings are consistent with prior studies [2, 7], including a large systematic review by Kulalert et al., where prolonged DPT nearly doubled diagnostic yield (Risk ratio 1.94) [8]. The mechanism underlying delayed positive challenge reactions is presently unclear. It is generally accepted that prolonged DPT aims to identify delayed hypersensitivity reactions; in our cohort, patients with positive delayed challenges were less likely to present with urticaria (11.4% vs. 30.0%, p = 0.043). However, we hypothesise that prolonged DPT can also identify immediate reactors who were re-sensitised, which could explain why some reactions only manifest after Day 3. Identifying these reactions is critical, as they might otherwise be misdiagnosed as non-allergic following a single-day DPT. We acknowledge that it is challenging to determine whether delayed reactions during prolonged DPT stem from prolonged antibiotic intake, or would have occurred following the in-clinic DPT dose alone. One potential solution is the use of a washout period; however, this may reduce compliance. Moreover, Chiriac et al.'s 20-year study of single-day DPTs for suspected beta-lactam allergies found that only 1.1% of positive provocations occurred after 48 h [9]. In our cohort, most positive prolonged DPT reactions (70.6%) emerged only from Day 3 onwards, suggesting these are true allergies that would have likely been missed without prolonged DPT. We also cannot exclude that a proportion of positive prolonged DPT reactions were due to re-sensitisation or de novo sensitisation. However, it is arguable that these reactions might have occurred irrespective of prolonged DPT if the patient had been prescribed a full penicillin course during a subsequent illness, which would be inconvenient and unpleasant. Additionally, some delayed reactions were self-reported, potentially contributing to the higher rates of delayed positive reactions observed. While patients were encouraged to submit photographs or return for in-clinic evaluation, some were unable or declined. The absence of post-challenge follow-up to confirm future penicillin tolerance is another limitation. Further studies are required to address these gaps. In conclusion, our findings suggest that a 5-day prolonged DPT may enhance diagnostic yield for suspected penicillin allergy in adults, while remaining safe. Patients with delayed onset or unknown index reactions may benefit from prolonged DPT. Further research is needed to determine the optimal duration and assess the utility of a washout period. H.Y.L. and K.J.L.C. contributed to the conception of the study. A.J.S.A. and C.J.C. were involved in data collection and verification. A.J.S.A. and K.J.L.C. were involved in data analysis. All authors contributed to the writing and/or proofreading of the manuscript. The authors declare no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request.