Discussion
This study included a large cohort of children referred to our center for BL allergy evaluation. We followed the classification of IR versus NIR, as initially reported by Terrados et al.10 and used extensively in the literature8,16,18,21-25.
Of the total patients evaluated, 10.6% were confirmed as allergic, most of whom were diagnosed by DPT. This proportion is lower than that reported by Ponvert et al.21, but consistent with other studies8,16,18,24.
Regarding the drugs involved, AX was implicated in 68.5% of the cases and AX-CLAV in 24%, very similar to results reported elsewhere16. Penicillin and cephalosporins accounted for 4.5% and 3% of the positive cases, respectively. Of interest is the important role of AX-CLAV, coherent with the high pattern of prescription1,2. The reported involvement of this formulation in DHR has changed from 12%8 to over 70% in recent years15,22,23. Considered less immunogenic than other BLs26, its contribution to both IR and NIR in adults has been reported27.
The proportion of confirmed IRs was 8.3%, with positive ST or SIgE contributing in the 5.4%. These values are lower than the 17% reported by Ibañez et al.16 but higher than the results observed by Mill et al.28 One study even reported a proportion of 86% positive ST21, but these data have not been supported by other studies8,16,23.
Regarding the different BLs, 63% of cases were selective reactors to AX, with no contribution from CLAV. This result contrasts with other studies published in adults, where 22% of participants were selective responders to CLAV27.
Although AX was involved in 92.5% of the cases initially evaluated, 27% of the confirmed responders belonged to the common group of penicillin reactor. The contribution of cephalosporins was less than 1%.
Symptoms suggestive of a NIR were reported in 377 cases, nearly 80% of the cases evaluated. Three cases with a severe reaction were diagnosed based on the clinical history. ST were positive in 1% of cases (2 cases to AX, 1 to PG and 1 to a cefotaxime). CLAV did not induce any positive ST. DPT was needed for confirmation in 36 of the 43 cases (83%) finally classified as positive. In total, considering all cases diagnosed based on clinical history, ST or DPT, the final proportion of positive NIRs was 11.4%. In our study, 52% of positive DPTs induced urticaria, with or without angioedema, while this reaction occurred in 72% of positive cases in the study by Ibañez et al.16 and 64% of cases included in Mori et al.’s report25. Data on the frequency of MPE versus NIU are variable in the literature. In large series like ours, the proportion of MPE reactions ranges anywhere from 18% to 80%15,18,22 which is in line with the 41.6% we observed in our study.
The interval between drug administration and symptoms onset was under an hour in IR, similar to what occurs in adults, although the dose required for eliciting a response was higher in children after correcting it for body weight29.
In NIRs, the analysis of the time required for inducing a positive DPT showed two response patterns. Up to a third of the cases were diagnosed on the first day, but over half the diagnoses required 6 or 7 days, which suggests that a 5-day protocol, as reported by some groups13,17,24, will miss an important number of positive cases. Other studies have also supported periods longer than 5 days26,27. In order to reconcile results from the different groups, responders with NIRs may fall into two broad groups, those responding early and those with a later response. This time interval is independent of the clinical entity induced (MPE or NIU), as shown in our work.
In NIRs, the role of selective responders to AX seems more relevant than in IRs, since 83% of cases with NIRs were selective in contrast with 63% of cases in IR. These data are similar to those reported in adults10.
One limitation of this study, as in other similar ones, is the lack of comparison between children showing a reaction at 2-3 days versus 5 days or longer. The low prevalence of allergy to BLs in children complicates this analysis. Also, we used patch testing, as done by other authors30,31 although there are studies suggesting that intradermal testing is more sensitive8,26.
Summarizing, after prospectively evaluating a large series of children with DHRs to BLs, we confirmed the diagnosis in 8.3% of the children assessed for IRs and 11,4% of those assessed for NIRs. Overall, NIRs were more frequent and had a higher proportion of selective responders. Moreover, reactions to CLAV were exclusively NIRs.