Validity of anamnestic data on allergies towards β-lactam antibiotics in a preoperative setting

R. Braun-Dullaeus Department of Cardiology and Angiology, University Hospital, Health Campus Immunology, Infectiology and Inflammation, Medical Faculty, Otto-von-Guericke-University Magdeburg, Magdeburg, Germany during preoperative examination. Using a self-designed questionnaire, we determined the timing of the allergic event, symptoms and previous diagnostics. Specific serum IgE (sIgE) were quantified (ThermoFisher [Waltham, MA, USA]; penicilloyl G and V, amoxicilloyl, ampicilloyl, cefaclor). Skin prick testing (SPT) was performedwith amoxicillin, cefuroxime, ampicillin, ciprofloxacin, moxifloxacin, benzyl-penicillin, piperacillin and meropenem in increasing concentrations (1:1000, 1:100, 1:10, 1:1). Immediate reactions were assessed as the largest diameter of wheal and flare and compared to controls. In all, N= 17 patients (16%) reported allergies to two or more AB (122 reports), most frequently to penicillin (n=81; 66%) and amoxicillin (n=13; 11%), while N= 57 patients (54%) reported >1 clinical manifestation. Most were delayed reactions (n= 58; 55%), often of a cutaneous type (n= 83; 52%). In 5 cases (3%), there was no information on symptoms. In 10 cases (6%), respiratory and in 11 cases (7%) cardiovascular symptoms were reported (Table 1). In 48% (n= 51), adverse reactions had occurred >10 years earlier. SPT revealed positive results in five cases. These had reported skin, respiratory and other reactions (Table 2). In two cases (1.8%), sIgE was >0.2kU/l (CAP I). In these, one reported a reaction >10 years ago and one a delayed type skin reaction with itching and hives. None of the SPT-positive patients showed increases sIgE. In the vast majority (93%; n= 99/106), the tentative diagnosis was not supported by the diagnostic results. Our study supports findings that an overwhelming majority of reported AB allergies is not reasonable. While valid documentation of adverse drug reactions should be performed as early as possible [5], it is often missing. Even in patients carrying an allergy pass, there was an overestimation of AB allergies. Most reactions were harmless, delayed skin reactions, indi-

SPT revealed positive results in five cases. These had reported skin, respiratory and other reactions ( Table 2). In two cases (1.8%), sIgE was > 0.2 kU/l (CAP I). In these, one reported a reaction > 10 years ago and one a delayed type skin reaction with itching and hives. None of the SPT-positive patients showed increases sIgE. In the vast majority (93%; n = 99/106), the tentative diagnosis was not supported by the diagnostic results.
Our study supports findings that an overwhelming majority of reported AB allergies is not reasonable. While valid documentation of adverse drug reactions should be performed as early as possible [5], it is often missing. Even in patients carrying an allergy pass, there was an overestimation of AB allergies. Most reactions were harmless, delayed skin reactions, indi- cating that they were non IgE-mediated maculopapular exanthemas [6,7]. Gastrointestinal symptoms are more likely to display non-allergic side effects, not confirmed as IgE-mediated allergies. Thereby, almost every tenth reported AB allergy could be refuted, as they exhibit gastrointestinal symptomatology [7] and there are promising attempts to discriminate true allergy, especially to β-lactam AB, based on medical history and symptoms [8].
In the majority of our patients, there was a long interval between the reported reaction and the study. This complicates documentation and may reduce the sensitivity of diagnostics [3,9]. Furthermore, sIgE is not necessarily relevant for AB allergy diagnostics [5]. Hypersensitivity to β-lactam AB can be lost, so that after 10 years 80% of patients tolerate penicillin [1,6,10]. However, a decrease in sIgE is not equivalent to the disappearance of allergic reactions [11,12]. To safely de-label reported allergies, an oral provocation test remains gold standard, while a well-structured anamnesis can be useful and less time-consuming [13].
Limitations of our study are that our methods only detect immediate IgE-mediated, but not delayed or pseudoallergic reactions. The sensitivity of SPT and sIgE is limited when judging type 1 allergic reactions to penicillins and type 4 allergy cannot be detected [14]. However, type 4 allergies are often mild and treatment can often be continued.
Overall, our results addressing IgE-mediated immediate allergies support the recommendation for diagnostics in questionable AB allergy especially in settings with planned, prophylactic administrations. There should be consequent efforts for de-labelling to support antibiotic stewardship [15,16].