LAUSR

Drug hypersensitivity reactions are an important public health concern and are frequently suspected in children, with a selfreported prevalence of up to 6% [1]. Non-immediate drug hypersensitivity reactions are defined as drug reactions occurringmore than 1 h after the administration of the drug dose [2]. Mild nonimmediate reactions generally consist of uncomplicated maculopapular exanthemas or delayed urticaria/angioedema and are devoid of danger signs [2,3]. Severe non-immediate reactions include severe cutaneous adverse reactions (SCARs) such as drug reaction with eosinophilia and systemic symptoms, StevensJohnson syndrome, toxic epidermal necrolysis, and acute generalized exanthematous pustulosis [3]. Beta-lactam (BL) allergy is the most common cause of drug allergy related consultations in children [3], with an incidence of 1–5% of skin eruptions per BL prescription [4]. Most rashes occurring during BL treatment are thought to be viral, as the majority are not reproduced with drug provocation tests (DPTs) [1]. However, in clinical practice, most of these patients are labeled as allergic without an appropriate allergic work-up to exclude or confirm a real allergy. Physicians and patients generally refuse to reuse the culprit drug once a suspected reaction has occurred, usually by fear of reproducing a similar ormore severe reaction [5]. Unfortunately, use of alternative antibiotics may lead to higher costs, treatment failures, and adverse reactions [6]. Traditionally, the evaluation of non-immediate mild reactions to BL involved delayed-reading intradermal tests (IDTs), followed by DPTs in patients with negative skin tests (STs). However, multiple studies have shown that IDTs have a low sensitivity [5,7–10] in non-immediate mild reactions to BLs. In addition, recent studies have found a low positive predictive value (PPV) of STs in both children [5,11] and populations consisting mainly of adults [12,13] when evaluating non-immediate mild reactions to BLs. Considering that these IDTs are not standardized, painful, and difficult to interpret, multiple centers have performed direct DPTs in children without skin testing before or without taking into account the result of STs [5,12,14–17], with no severe reactions being documented. Nonetheless, this approach is still controversial, and some centers [7] continue to perform STs in mild non-immediate reactions in children to reduce the risk of having even one positive DPT, arguing that patients with positive IDTs have a higher rate of positive DPTs [18]. From another point of view, a large number of painful IDTs need to be performed to detect only a very small amount of positive reactions on DPT, which have been shown to be mild in numerous studies [5,11,12]. Furthermore, in our experience, many parents refuse painful IDTs to be performed on their childrenwhereas the DPT is well accepted [19]. Therefore, direct DPT without STs in children with mild non-immediate reactions is increasingly accepted, and this approach has recently been outlined by the European Academy of Allergy and Clinical Immunology [3] and British Society of Allergy and Clinical Immunology [20]. However, additional multicentric studies are still required to validate the safety of such an approach and increase overall consensus. Thus, DPTs play a central role in the management of children with a suspicion of non-immediate BL allergy. However, DPTs are not standardized and multiple different protocols are used among centers, even in the same country. In particular, the optimal length of DPTs remains highly debated, ranging from 1 to 10 days [5,12,14,17,21–23]. It has been shown that longer protocols can increase the sensitivity, with up to 6.2% of patients reacting after the first day of DPT [5,7–9,14,17,21]. In addition, some authors have suggested that compliance to BLs may be increased after extended DPTs when compared to shorter protocols [14,23]. On the other hand, the negative predictive value (NPV) of 1-day DPTs is relatively similar to the NPV of prolonged DPTs (89.1 to 97.6%) [5,14,22,24–26]. Of note, the percentage of patients that will react after a negative DPT is similar to the incidence of cutaneous rashes during antibiotic treatments in children, which is estimated around 5% [4]. In addition, an important concept is the number needed to harm [25]; many children need to be exposed to a prolonged course of antibiotics to screen a few patients whowould present amild reactionwhen retaking the antibiotic. Although this may seem innocuous, taking antibiotics for no validated reason is not benign; studies have shown that gut microbiota is significantly altered during antibiotic course [27], and long-term effects of multiple antibiotic courses in children are numerous, including obesity [28], antimicrobial resistance [29], and possibly increased atopy [30]. Another possible confounding factor in studies is the delay between the index reaction and the DPT. The delay is variable, with some studies having a delay of only a few months [8], and others a mean of 2 to 3 years [14,24]. When the delay is longer, parents often cannot recall the reaction, and it is often difficult to classify the patients as immediate or non-immediate reactions. In