Ara h 2 and Ara h 6 sIgE determined using ISAC analysis are good predictors for peanut allergy and ISAC cut-off values are reliable and useful to predict a clinically relevant peanut allergy. Ara h 2 sIgE determined by multiplex ISAC microarray analysis shows a good correlation with Ara h 2 sIgE determined by singleplex FEIA measurement. To our knowledge, this is one of the first studies evaluating cut-off points of peanut components using multiplex ISAC analyses in peanut allergic and peanut tolerant children, diagnosed by double blind placebo controlled peanut challenge tests. The few other studies that have investigated sIgE cut-off values to diagnose a peanut allergy using microarray peanut components in children did not use food challenge tests as golden standard or performed peanut challenge tests in only a subgroup of the study population. [14, 23, 24, 29, 30]
In our population Ara h 2 and Ara h 6 scored best in diagnosing peanut allergy. 100% PPV was reached at ISAC microarray cut off values of 7.43 ISU and 8.13 ISU for Ara h 2 and Ara h 6 sIgE respectively. Nineteen and 24 patients (16% and 21% of the study population) had an Ara h 2 and Ara h 6 value above this cut off value and could be classified as peanut allergic with 100% certainty. Klemans et al. calculated cut off values in diagnosing peanut allergy in adults using ISAC, and showed a 100% PPV at a cut off value of >9.74 and >2.40 for respectively Ara h 2 and Ara h 6. [31] Looking at our population, 2 outliers increased the cut off values of Ara h 2 and Ara h 6 from 2.47 and 2.32 ISU to 7.43 and 8.13 ISU respectively. Both patients passed the 6-dose food challenge. However, in one patient information regarding home introduction was missing due to loss to follow-up and in the other patient, home introduction failed due to refusal by the child.
The few other studies that evaluated the performance of peanut components in diagnosing peanut allergy in children using ISAC analyses showed similar results revealing Ara h 2 and Ara h 6 as best performing allergens. Cut off values in these studies vary most probably due to other study designs, study population and geographical location.
Singleplex versus multiplex Ara h 2 analysis
The semi-quantitative sIgE measurements of multiplex microarrays are considered to be less sensitive for monitoring sensitization compared to singleplex measurements. [32] In our study, Ara h 2 sIgE values in singleplex FEIA and multiplex ISAC showed a strong correlation and we did not miss more peanut allergies using ISAC compared to FEIA Arah2 sIgE. In addition, the simultaneous measurement of other peanut components, Ara h 6 sIgE in specific, may improve the diagnostic performance of ISAC microarray. This is in line with other studies, which also showed comparable IgE recognition patterns and diagnostic sensitivities between multiple and singleplex determined peanut allergens. [33-35]
Cut-off values of singleplex Ara h 2
Our singleplex Ara h 2 sIgE data confirmed sIgE to Ara h 2 as a good predictor for peanut allergy in children. [8, 17, 20, 36-45]. In our study we found a 100% positive predictive value for Ara h 2 sIgE determined by singleplex FEIA at 4.40 kU/l. Cut-off values can vary due to study population and geographical location. To use cut off values in daily practice, clinicians has to determine and validate cut off values in their own specific population and region. Nevertheless, our results are quite similar to two other Dutch studies in comparable atopic pediatric study populations but in other regions of the country. [36, 46]
In the study from Klemans et al. the negative predictive value of FEIA Ara h 2 sIgE improved to 100% accuracy lowering the lower cut-off value from 0.35 kU/l to 0.07 kU/l. They concluded that the need for peanut challenges could be reduced using Ara h 2 sIgE measurements. In our study population 5 patients (8%) of the Ara h 2 sIgE negative children, determined by FEIA developed objective allergic symptoms upon peanut challenge. This is in concordance with a recent systematic review based on 16 control studies in children showing that the Ara h 2 sIgE cut-off value of 0.35 kU/l results in 8.1% false negative results. [21] Based on these findings we advise to consider a peanut home introduction carefully in Ara h 2 negative children due to the small risk of an allergic reaction at home.
The small proportion of peanut allergies in children with very low or negative sIgE to Ara h2 and whole peanut extract values can be explained by either a lower threshold of sIgE to Ara h2 provoking allergic reactions or sensitization to other peanut minor components such as oleosins. Oleosins are lipophilic allergens that are underrepresented in whole peanut extracts because they are poorly soluble in aqueous solutions. [47]
Peanut component patterns
We evaluated sensitization patterns of different peanut components in peanut allergic and peanut tolerant children. Our data showed that Ara h 2 and Ara h 6 are the most common peanut allergens in children with a peanut allergy (94% and 92%), followed by Ara h 1 and Ara h 8 detected in 49% and 38%.
This is similar to the 90% prevalence of Ara h 2 sensitization in peanut allergic children reported by others [30, 48, 49], but in contrast with other studies, that reported a lower prevalence [24, 29]. Variations in reported prevalence of sIgE to peanut components may be linked to differences in study designs, geographical location and study populations.
We found, in line with other studies, a high rate of co-sensitization for Ara h 2 and Ara h 6 in peanut allergic children. [24, 29, 31, 37, 50-52] Ara h 2 and Ara h 6 are both seed storage proteins belonging to the 2S albumin family and they share a high amino acid sequence identity. However, mono-sensitizations to Ara h 6 occur in peanut allergic children. In our study we detected two peanut allergic child sensitized to Ara h 6, but not to Ara h 2. This enforces the added value of Ara h 6 sIgE determination in Ara h 2 negative children with a possible peanut allergy. This is similar to some other studies, who detected mono sensitizations of Ara h 6 in 1.2-18% of the peanut allergic children and adults [7, 24, 31, 49], but is in contrast with others that did not detect any Ara h 6 mono sensitization. [33, 53]
Although we detected Ara h1 and Ara h3 in respectively 49% and 27% of the peanut allergic children, these storage proteins did not contribute to the diagnosis of peanut allergy in our population since no peanut-allergic patients were sensitized to Ara h 1 or Ara h 3 without co-sensitization to either Ara h 2 or Ara h 6.
Also other studies show a relatively high prevalence of sensitization to Ara h1 (between 40% and 94%) and Ara h 3 (between 23% and 77% ), in peanut allergic children and adults, but only few mono-sensitizations to Ara h 1 or Ara h 3 are found in the peanut allergic population. [24, 29, 33, 37, 49]
Sensitization to Ara h 8 was frequent in both the peanut allergic and peanut tolerant group, but not significantly different between both groups. The high prevalence of sIgE to the birch pollen homologue Ara h 8 is consistent with the high prevalence of birch pollen sensitization in our Dutch population comparable with other North European studies. [17, 19, 49, 54] Sensitization to the nsLTP component Ara h 9 was low in our population and did not contribute to the diagnostic accuracy peanut component analysis. This is in line with other reports from northern European populations [19, 48].
In earlier studies, sensitization to multiple peanut storage proteins has been associated with a higher probability of clinically relevant peanut allergy. [17, 37, 49, 54] Our study shows similar results, 91% of peanut allergic children were sensitized to at least 2 storage proteins compared to 22% in the children with a negative peanut challenge. Mono-sensitization to storage proteins was less frequently found in the peanut allergic group.
There are some limitations of this study. In the peanut allergic group, children more often presented with allergic reactions to peanut in the history compared to the peanut tolerant group. This could have biased the results. However, another study showed no differences in specific IgE to peanut and major peanut allergens between challenge confirmed peanut allergic children with a clinical history to peanut compared to them with sensitization without earlier exposure to peanut. [49]
Another limitation of this study was that 2 different peanut challenge test dosing schemes were used over time and between the two participating centers before 2014. This may have influenced the calculated sensitivity of the different sIgE tests and the associated negative predictive value. However, most patients with a negative peanut challenge succeeded in home introduction of the peanut. In 1 patient with a negative peanut challenge but increased Ara h 2 and Ara h 6 sIgE information regarding home introduction is missing due to loss to follow-up and 1 patient did not succeed to introduce peanut at home because of refusal.