DISCUSSION
In the present study, we found that most of children with AD were sensitised to at least 1 food allergens, and those who were sensitised to a greater number of FAs had more severe AD phenotypes and greater epidermal integrity impairment. Moreover, SCORAD score and SCH level could predict FS in children with AD.
Most children with AD (90%) in the present study were sensitised to at least one of the tested FAs; this prevalence was higher than those reported by Moghtaderi et al.8 (51%) and Yuenyongviwat et al.9 (60%). However, those studies investigated smaller numbers of FAs (20 and 8 allergens, respectively), compared with our study.8,9 Additionally, other studies in children aged < 15 years showed that the prevalence of FS ranged from 30% to 40%,10,11 below the prevalence in our study. This difference may have occurred because FS is more frequent in young children and can be outgrown. In another study, the prevalence of FS in children with AD ranged from 20% to 80%.12 These findings suggest that FS is more common than previously reported in children with AD, particularly among such children aged < 5 years. Additionally, the number of FAs investigated in each study and the ethnicity of the study population could affect FS prevalence.
We found that cow’s milk protein (60%) was the most common sensitised FA, followed by egg white (49%), beef (31%), almond (27%), egg yolk (26%), and peanut (26%). Similar results were reported by Moghtaderi et al.8. Another study also showed that egg, cow’s milk, and peanut were the most prevalent allergens in children with AD who were sensitised to FA.13 However, a study in an American population showed that egg and peanut were the most prevalent FAs.14 Differences in common sensitised FAs among studies could be related to differences in ethnicity, age, local food customs, and dietary habits among study populations. In Vietnam and other Asian countries, cow’s milk proteins and egg appear to be the most common FAs in children.
We also examined features of co-sensitisation among the tested FAs and found that egg white and beef had the highest prevalence of co-sensitisation, such that 71.4% of children sensitised to egg white were also sensitised to beef; 58.3% of children sensitised to beef were also sensitised to egg white. This prevalence was higher than that reported previously, which showed that 28.5% of AD children who were sensitised to egg white also exhibited beef sensitisation.15 The prevalence of co-sensitisation to cow’s milk and cow’s milk protein ranged from 75% to 91.7% in children with AD who were sensitised to goat’s milk. Therefore, in clinical practice, caution should be exercised when introducing cow’s milk to AD children who have had allergic reactions to beef and/or goat’s milk. Additionally, children with AD who were sensitised to egg yolk demonstrated a high prevalence (92.3%) of sensitisation to egg white; consequently, caution should be exercised when introducing egg white to children with AD who have had allergic reactions to egg yolk.
To our knowledge, this is the first study in Vietnam to evaluate the association between FS and epidermal layer impairment. We found that in children with AD, lesional skin had significantly higher TEWL levels and lower SCH levels compared with non-lesional skin. Additionally, participants with moderate-severe AD had significantly higher TEWL levels and lower SCH levels in lesional skin, compared with participants who had mild AD. These findings were consistent with the results of Montero-Vilchez et al.16, who showed that impaired skin barrier function in the lesional skin was associated with AD severity.
The damaged epidermal barrier (indicated by increased TEWL levels and decreased SCH levels) in patients with AD could enhance FA penetration and induce FS.2 In the present study, participants who were sensitised to more FAs exhibited higher TEWL levels and lower SCH levels, compared with non-sensitisers or children sensitised to fewer FAs. A previous study involving children aged 1–2 years with AD and FS revealed higher TEWL levels than non-sensitisers.14Although these findings suggest a clear association between FS and epidermal barrier impairment, the causative relationship between these factors cannot be determined using data from cross-sectional studies.
We found that the number of sensitised FAs was positively associated with AD severity, as determined by SCORAD scores. Compared with non-sensitisers, children with AD who were sensitised to FAs such as egg, cow’s and goat’s milk proteins, beef, pork, chicken, rice, and others (Table S1) had more severe AD. This was consistent with the findings of Wolkerstorfer et al.17. Additionally, Leung et al.18 showed that decreased expression of filaggrin and increased TEWL levels in the skin of individuals with AD could facilitate FS into deeper skin layers. Subsequent exposure to FAs via damaged skin could trigger an inflammatory response, leading to enhancement of AD severity.2 Other studies showed strong associations among FS, food allergies, and the severity and chronicity of AD.8,18 Consequently, those findings indicate that FS could be associated with epidermal barrier impairment and subsequently AD severity in children.
There is currently no clinical guidance regarding the appropriate time to evaluate FA in patients with AD. The results of previous studies have suggested that FS is associated with severe and persistent AD.19 In the present study, we analysed the ROC curves of SCORAD scores, as well as the utilities of SCH and TEWL levels, in terms of predicting FS in the study subjects. We found that SCORAD scores and SCH levels in lesional skin had moderate predictive value for sensitisation to cow’s milk, egg, almond, beef, peanut, and goat’s milk in children with AD. SCORAD scores and SCH levels in lesional skin had the highest predictive value for sensitisation to peanut and the lowest predictive value for sensitisation to cow’s milk. However, TEWL levels were not useful for predicting FS in the present study. The SCORAD assessment and the measurements of TEWL and SCH levels are easy, non-invasive procedures that are safe for children. Our findings suggest that clinicians should consider assessing FS in children with moderate-severe AD and/or low SCH levels in lesional skin.
The limitations of the present study include the inability to perform a follow-up assessment because of the cross-sectional design. Furthermore, we did not perform an oral food challenge test to determine food allergies. The age range of our study participants was also limited because we did not include children aged < 1 year. Thus, the study results are not representative of children with the highest prevalence of AD (i.e., children aged < 6 months).
In conclusion, FS was common in children with AD; it was closely associated with AD severity and epidermal barrier impairment. Evaluations of food sensitisation should be considered for children with moderate-severe AD and/or low SCH levels.