Beta-lactam allergy is a common problem in everyday medical practice and is recognized as a major public health issue (1). Studies indicate that ~10% (2) of the general population and 15% of the hospital population (3) are labeled with a penicillin allergy. This label is often misleading because contemporary publications indicate that <10% of these patients are truly allergic after an oral drug challenge (4). Carrying this label frequently leads to the avoidance of all beta-lactam antibiotics, favoring the use of other less preferred classes of antibiotics, that are more expensive and associated with more side effects and increased antimicrobial resistance (5). Delabeling a beta-lactam allergy is part of antimicrobial stewardship programs.

Since 2011, a decisional algorithm has been set up at the Centre Hospitalier Universitaire Vaudois (CHUV) to optimize the management of patients labeled with a beta-lactam allergy (Figure 1). This algorithm is based on clinical history, skin testing [prick tests and intradermal reactions (IDR)] and drug challenges (oral, intravenous, or intramuscular) (6) and allows to identify patients who are truly allergic and to delabel the allergy to beta-lactam in whom it is not proven.

In this study, we performed a retrospective analysis of patients who were referred to our university allergy clinic with a history of allergy to beta-lactam antibiotics. We assessed the true frequency of immediate and delayed allergies to commonly prescribed subclasses of beta-lactam antibiotics (including penicillins and cephalosporins) and the negative predictive value (NPV) and the sensitivity of skin tests. Our secondary aims were to examine the safety of beta-lactam skin testing and drug challenge.

Medical files of all consecutive patients seen at the outpatient clinic between January 1st 2011 and December 31st 2018 were retrospectively screened for the following keywords: “beta-lactam, penicillin, cephalosporin, carbapenem, amoxicillin, amoxicillin/clavulanate, flucloxacillin, piperacillin-tazobactam, cefuroxime, cefpodoxime, cefazolin, cefepime, imipenem-cilastatin, ertapenem, meropenem, penicilloyl-polylysine (PPL), minor determinant mixture (MDM).” Any patients with a beta-lactam antibiotic listed in their electronic files were included even if there was no notion of an allergy (for example, the absence of an allergy to beta-lactam antibiotics was often listed in the medical file and was detected by the keyword screening). Further examination of the medical records that met the above criteria identified patients who were referred to investigate a history of an allergy to beta-lactam antibiotics. Of note, patients with an allergy to carbapenems were removed from the analysis due to their limited number.

The following data were collected for patients referred for a history of an allergy to beta-lactam antibiotics: age, gender, history of atopy and autoimmune disease, history of reaction to beta-lactam antibiotics (type, grade, date, antibiotic molecule involved, and medical management of the reaction), beta-lactam skin test results, beta-lactam drug challenge results, adverse reaction to skin tests and drug challenges.

Immediate reactions were defined as reactions that occurred up to 6 h but usually during the first hour after beta-lactam exposure and included the following clinical features: hives, angioedema, bronchospasm and/or hypotension (7). Severity of the reaction was graded from I to IV according to Müller's grading system (8) (in brief, grade I: hives; grade II: angioedema, gastrointestinal manifestations; grade III: bronchospasm, grade IV: hypotension). Delayed reactions were defined as reactions that occurred more than 6 h after drug exposure and featured maculopapular eruptions or more severe hypersensitivity reactions such as drug reaction with eosinophilia and systemic symptoms (DRESS), Stevens-Johnson syndrome (SJS), and toxic epidermal necrolysis (TEN). For reactions occurring between 1 and 6 h, we used the phenotype of the reaction at first to determine if it was an immediate or a delayed-type of reaction. Unknown timing or mechanism of reactions were defined by reactions for which information on timing or the mechanism of reaction after beta-lactam exposure were missing.

A diagnosis of confirmed allergy was based on a positive skin test and/or drug challenge. Skin testing workup is summarized in Supplementary Table 1. If the phenotype of the reaction was unknown, skin testing should include both immediate and delayed readings.

The performance of skin testing was evaluated in patients with results of both skin test and drug challenge, with distinction between immediate and delayed-type reactions. In rare instance, a direct drug challenge (i.e., without prior skin testing) was performed on the basis of the physician's assessment because the risk of a reaction was highly unlikely. The majority of drug challenges (N = 270) consisted of a single drug dose. A minority of drug challenges (N = 18) were realized on 3 days with a single drug dose per day according to the decision of the investigator.

We examined the safety of the current standardized approach to evaluate patients with a beta-lactam allergy label, which includes a thorough evaluation of the medical history, followed by skin tests and drug challenge.

The study was conducted in accordance with the Declaration of Helsinki. The project was approved by the Ethics Committee of the Canton of Vaud (CER-VD 2019-00337) that waived the need for informed consent allowing the inclusion of all patients except those who refused the use of their data. All data were anonymized before analysis. Data analysis was performed using Stata software version 16.1 for Windows (Stata Corp LCC).

Of the 2,320 patients screened for a history of allergy to beta-lactam, 576 met the inclusion criteria and were included in the analysis (Tables 1, 2). Mean age of the patients was 49.7 years (SD ± 17.7) and 65.3% were female. Eighty five patients reported more than one reaction (which included both multiple reactions to the same class of beta-lactam antibiotics and reactions to different classes of beta-lactam antibiotics). Patients were divided into three groups according to the reported type of reaction: 309 patients had a history of an immediate-type and 172 of a delayed-type reaction (Table 1). In 138 patients, the timing or mechanism of the reaction was unknown (Table 2). Twelve patients reported reactions to both penicillins and cephalosporins in the immediate-type reaction group, and 8 in the delayed-type reaction group. Twelve patients reported both immediate- and delayed-type reactions and were included in both groups.

In this study, we evaluated 576 patients who were referred to our university hospital allergy clinic for a history of allergy to one of the beta-lactam antibiotics. Overall, 678 reactions were described.

Since immediate- and delayed-type allergies have different pathophysiological mechanisms, patients were separated into distinct groups according to the clinical history of their reactions: immediate (309/678 = 45.6%), delayed (172/678 = 25.4%) or unknown timing or mechanism of the reaction (138/678 = 20.4%). Each group had similar demographic characteristics. Interestingly, there is a significant preponderance of women in each patient group, with a female to male ratio of ~2:1. This aspect is in accordance with previously published studies (9, 10). Review of the clinical history indicated that the most commonly suspected drugs were penicillins (544/678 = 80.2%) followed by cephalosporins (127/678 = 18.7%). This distribution primarily reflects the frequency of beta-lactam subclasses prescription.

Our main aim was to assess the prevalence of confirmed beta-lactam allergies in the study population. Considering global beta-lactam allergy, allergy was confirmed in 13.4% of the patients with a history of immediate-type reaction, while delayed-type allergy was confirmed in only 4.9% of the patients. Overall, the prevalence of confirmed allergies to beta-lactam antibiotics based on positive skin tests and drug challenges was 18.3%.

Deeper examination of subclasses of beta-lactam antibiotics highlighted that confirmed penicillins immediate-type allergies occurred in 10.9% of the patients and confirmed delayed-type allergies in 5.6%. The results, particularly for immediate-type allergy, closely match previous published data, as most of the studies show that around 90% of the patients who are labeled with penicillins allergy tolerate this drug (5, 11, 12).

Regarding cephalosporins allergy, 18% had a confirmed immediate-type allergy, while only 0.9% of patients were identified with an actual delayed-type allergy. The data available on the epidemiology of cephalosporins allergy are more limited than those on penicillins allergy (13–17). A recent large French retrospective study found that allergy to cephalosporins was proven in 22.3% of the patients and the majority of these reactions were immediate-type (18). According to our data, patients who report a history of allergy to cephalosporins are more likely to be allergic than patients who report an allergy to penicillins (overall 18.9% vs. 16.4% of patients). Furthermore, we confirm that a delayed-type allergy to cephalosporins is a rare condition, as already shown by others (13, 18).

We found that NPV of skin tests for penicillin antibiotics were 96.3% and 91.9% for immediate- and delayed-type reactions, respectively, meaning that patients with negative skin tests will unlikely present an allergic reaction on drug challenge. Sensitivity of skin tests for penicillins was 90.7% for immediate-type reactions and 84.2% for delayed-type ones. These values are similar to data previously reported in retrospective and prospective studies (19, 20). For immediate- and delayed-type reactions to cephalosporins, we determined a NPV of skin tests of 100 and 87.5%, respectively. These results are based on 17 patients for immediate-type, and 8 patients for delayed-type reactions, which likely overestimate these values compared to recent larger studies (18).

Safety of skin tests showed that only 0.7% of the patients presented a systemic allergic reaction. Importantly, only one patient developed a malaise with a transient hypotension. All patients requested antihistamine treatment and for two of them (one for bronchospasm and one for generalized skin hives) the inhalation of a bronchodilator. Beta-lactam skin testing is regarded to be a safe procedure if performed by a trained medical team who is prepared to treat anaphylaxis (5, 11, 21). Assessment of the safety of beta-lactam drug challenges indicated a rate of systemic allergic reactions of 3.1%. These data are similar to previously published results (11, 12, 22–28). Careful examination of these systemic allergic responses showed that most of the reactions were benign skin rashes (hives or maculopapular skin eruptions) but two anaphylaxes were reported and, interestingly, two type-III reactions (purpuric skin rash, arthralgia and fever). These retrospective data emphasize that beta-lactam drug challenge is globally a safe procedure, which should take place in a medical care structure prepared to recognize and treat drug hypersensitivity.

The limitations of this research include the low number of patients in some subgroups. Accordingly, results of NPV for cephalosporins are based on only 17 patients for immediate-type reactions, and 8 patients for delayed-type reactions, which likely overestimate these values compared to recent larger studies (18).

Of note, there are no clear guidelines for defining the duration of the drug challenge for delayed-type reaction. Therefore, in this real-life study, delayed challenges were performed either with a single drug dose (N = 36) or with consecutive administration of the drug for 3 days (N = 17). Our analysis of the NPV and skin tests sensitivity for delayed-type reactions did not take this variability into account. Therefore, our results were influenced by the duration of the challenges, and the conclusions of these results should be taken with some caution.

In conclusion, in this study we examined patients with a history of allergy to beta-lactam antibiotics under real-life conditions. We have shown that a complete allergy workup, including careful evaluation of the medical history, skin tests and drug challenge, helps to delabel patients who are not truly allergic, and who represent the vast majority of patients included in the present study. This approach contributes to save second line antibiotics that are associated with more side effects and increased antimicrobial resistance. In addition, a standardized allergy workup is a safe procedure if performed by a trained medical team ready to identify and treat allergy and anaphylaxis.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee of the Canton of Vaud, Switzerland (CER-VD 2019-00337). The ethics committee waived the requirement of written informed consent for participation.

DC: conception and design of the study. SI: acquisition and analysis of the data. SI, LS, LM, YM, CR, GB, and DC: interpretation of data. SI and DC: drafting the manuscript. LS, LM, YM, CR, and GB: revising the manuscript. SI, LS, LM, YM, CR, and DC: final approval of the version to be published and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All authors contributed to the article and approved the submitted version.

Open access funding provided by University of Lausanne.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.