DISCUSSION
In this retrospective Portuguese pediatric study that included 98 patients, the most frequent nuts involved were peanut (63%), hazelnut (59%) and walnut (53%). Mean age at first reaction was 5.8 years, being younger in patients with peanut allergy comparing with other nuts (4.6 years versus 6.7 years, p=0.02). We assume that it is a possible consequence of a recent rising nuts consumption in pediatric population in our country, predominantly of peanut. In Spain, Haroun-Díaz Eet al . reported these same three nuts (hazelnut, peanut and walnut) as the most frequent nuts eliciting allergy.15Allergy prevalence for each TN seems to vary in different regions of the world: hazelnut allergy is the most frequent in continental Europe; peanut, brazil nut, walnut and almond are the most commonly reported in the United Kingdom, and walnut and cashew allergies in the United States. These differences are representative of the variations in nuts consumption in each country, leading to different sensitization patterns.8, 16,17
Most patients had history of atopy (n=86, 88%), including 44% with eczema and 31% with egg allergy. Cetinkaya PG et al. reported a higher frequency of these atopic conditions in patients allergic to nuts, with 72% having atopic dermatitis and 50% egg allergy, probably related to a higher number of involved children.18 The epithelial barrier dysfunction characteristic of atopic dermatitis is a confirmed risk factor for the development of allergic sensitization, food allergy and other allergic diseases. Many epidemiological studies, and recently also studies in animals, demonstrated the connection between skin and digestive tract. Damaged keratinocytes produce IL-33, that stimulates group 2 innate lymphoid cells (ILC2) in the small intestine. These in turn produce IL-4 and IL-13, which leads to the expansion of activated mast cells, resulting in an increment of intestinal permeability and consequent transmission of allergens that can trigger food allergy.19,20
Nut reactions may be severe on the first contact. About one third of our patients presented with anaphylaxis, and more than half were polysensitized to several nuts, which in accordance with recent data of international cohorts.1, 18 Avoidance of all nuts has been the rule for many years in patients allergic to one nut, but the possible introduction of other nuts has recently been investigated in several studies.21,22 In our study, 9 out of the 15 OFCs performed with non-culprit nuts to which patients were sensitized, but not exposed to before, were negative, favoring the above-mentioned idea of avoiding unnecessary nuts restrictions. Additionally, 31 (32%) were monosensitized after allergological investigation (16 to peanut, 7 to walnut, 4 to cashew, 2 to hazelnut and 2 to pine nut).
IgE sensitization to pan-allergen LTP were the most prevalent in our population, differently from a Spanish cohort that found the 2S albumin family of the seed storage proteins as the most frequent.1 It is, however, in line with recently published Portuguese data.23
In our pediatric cohort, anaphylaxis was more common in patients with chestnut and cashew allergies. Severe systemic reactions in patients sensitized to cashew have been frequently reported in Europe.9,12,14 A Portuguese study on TN anaphylaxis in preschool age children concluded that cashew was the major culprit, accounting for 11 of the 25 cases.24 Data on chestnut allergy is sparce, especially considering prevalence and reaction’s severity. Our results could be explained by Portuguese eating habits, with chestnut being one of the most appreciated nuts, typical of the colder months.
Allergological investigation proved to be useful in anaphylaxis risk prediction. MPD should be valued not only for diagnosis but also for anaphylaxis risk prediction. In our study, MPD was significantly higher for almond, cashew and pistachio in patients with anaphylaxis to these TNs. Other authors reported utility of ST in diagnosis prediction (MPD ≥ 8mm), but not in anaphylaxis risk prediction.9,12 sIgE did not differ significantly between groups. However, it should be taken into account that there was limited availability of whole extract sIgE for some of the nuts analyzed (namely cashew, pistachio, chestnut and pine nut) and that not all patients did sIgE measurements. In addition, as previously postulated,25,26 our study highlights the important adding value that sIgE/total IgE ratio could bring to clinical practice. For example, sIgE measurement for walnut was not a good predictor of anaphylaxis but, when integrated in sIgE/total IgE ratio, it reached statistical significance. Component-resolved diagnosis was used as a study complement, showing interest in risk stratification, particularly for peanut, with Ara h 2 and Ara h 6 being significantly higher in patients with anaphylaxis, which is in line with other reports.26
The present study has some limitations, such as the small sample of patients, which limits extrapolation of the results, and its retrospective design, that could itself weaken our findings. However, it is the first Portuguese study which extensively characterizes a pediatric population with TN and peanut allergy, analyzing the clinical utility of ST, sIgE and mcIgE measurements and, importantly and sparsely reported, sIgE/total IgE ratio in anaphylaxis prediction, and so the authors believe that it could add significant value to clinical practice.