The combination of Ara h 2‐sIgE and basophil activation test could be an alternative to oral food challenge in cases of suspected peanut allergy

Most children with peanut sensitisation do not have a clinical peanut allergy (PA). Oral food challenge (OFC) is then necessary to diagnose PA and assess the reactive dose of the allergen. However, OFC is laborious to perform, expensive and stressful. We evaluated whether in vitro tests, such as basophil activation test (BAT), allergen‐specific IgE (sIgE) and their combination, could be used to replace OFC for the diagnosis of PA in children.

diets are particularly restrictive for children and their families because of the ubiquity of peanuts or small traces of them in our daily diet. 3 Because food allergy has a significant impact on patients' life and unnecessary dietary restrictions can be harmful, an accurate diagnosis is of utmost importance. 4 While the prevalence of peanut sensitisation varies between 2.3% and 10.1% in Europe, only 0.3%-0.9% of these patients are clinically allergic, implying that testing is needed to distinguish these patients. 5 The oral food challenge (OFC) remains the gold standard for food allergy diagnosis. 6 Even if patients have strong evidence of PA (clear association between peanut consumption and allergic manifestations along with sensitisation), it is generally mandatory to assess the allergen reactive dose in order to design an OTI protocol. However, this test remains time-consuming, expensive, stressful and sometimes difficult to interpret, especially in young children. 7 In this context, reliable biological markers of clinical allergic reaction were investigated. Thus, the detection of allergen-specific IgE (sIgE) and molecular allergens such as Ara h 2 was shown to discriminate between patients with PA and patients sensitised to whole peanut extract but not clinically responsive to peanut. [8][9][10] More recently, the basophil activation test (BAT) has emerged as an interesting in vitro functional assay that goes even further than just exploring sensitisation. It measures the expression, by flow cytometry, of surface activation markers such as CD63, a highly specific marker for basophil degranulation. 8,9 Santos et al. suggested that BAT may be regarded as an 'OFC in a test tube', becoming an alternative for standard procedures and their inconveniences. 9 Thus, in some studies, BAT proved to be superior to other diagnostic tests in discriminating between peanut allergy and tolerance. 9,11 BAT was also able to estimate the severity and threshold of allergic reaction in PA. 10 However, assessments of BAT in food allergy remain scarce and further studies are needed to define and validate diagnostic cutoff values and to standardize the methodology to allow a wider use of BAT in clinical practice. 12 In this study, we aimed to assess the diagnostic performance of BAT in children with suspected PA and to evaluate whether a combination of BAT and other in vitro assays (Ara h 2-or whole peanut-sIgE) could be used to replace OFC.

| Study design
Patients from 6 months to 18 years followed at our tertiary care centre for a suspected PA were prospectively recruited between

| Oral food challenge
Open-label OFCs were performed in accordance with local clinical protocol. Placebo was given as the first dose then peanut oil was tested as a nonallergenic peanut product to reassure parents on the consumption of peanut oil. Thereafter, fresh homogenised peanuts with amounts of peanut proteins equal to 12.5, 25, 37.5, 62.5, 125, 500, 1000 and 2000 mg were given orally every 30 min. The test was monitored, and clinical reactions were evaluated. The decision to discontinue the OFC was made when objective symptoms occurred.
Objective symptoms were defined as clinical manifestations scoring from 1 to 5 in the Astier's anaphylaxis score. 13 Thus, the OFC was

| Skin prick test
Peanut SPT was performed using commercial extracts and con- We demonstrated the good diagnosis performance of the basophil activation test. Furthermore, we propose a diagnostic algorithm in order to use this test to discriminate nonallergic patient and significantly reduce the need for oral food challenge.
wheal diameters (mm) were measured after 15 min, and when then the wheal diameter was ≥3 mm, the SPT was considered to be positive.

| sIgE evaluation
All available serum samples were analysed using ImmunoCAP 250 (Thermo Fischer) for sIgE reactivity to whole peanut, Ara h 2 and Ara h 8.

| Basophil activation tests
Heparinised whole blood samples were collected in the hospital, just before undergoing OFC, and processed within 4-24 h. Whole

| Statistical analysis
The principal aim was to estimate the sensitivity of BAT, with OFC as the gold standard. For this purpose, the confidence interval (CI) for sensitivity was constructed according to Clopper-Pearson's method.
In the same way, the CI for specificity was established. Population characteristics were described using absolute numbers (N) if they were nominal, otherwise the median as well as the first and third quartiles (Q1-Q3) was used.
Logistic regression was used to assess the association between characteristics and the OFC outcome. The odds ratio (OR) for continuous variables indicates the odds per unit change. As this analysis is exploratory, a p-value <.05 is considered to indicate a significant association.

| Study population
Ninety-one children and adolescents with a suspected PA were prospectively recruited ( Table 1). All of them were included because they all had a positive SPT. Median age was 6.8 years (Q1-Q3: 5.0-11.3 years), and 64% were boys. Most patients had asthma or rhinitis (74%), which were objectified by upper and lower respiratory signs such as bursts, nasal pruritus, persistent rhinorrhea, expiratory wheezing to auscultation or audible wheezing. Atopic dermatitis was also frequent (60%). Poly-allergy, defined as allergy of at least two food allergens, was found in 45% of patients. Except for one, all patients were on a peanut exclusion diet before OFC. This patient reconsumed peanuts accidentally and had no reaction.

| Oral food challenge
Peanut OFCs were positive in 41 out of 91 patients (45%). Most of these patients developed urticaria or angioedema (71%), 27% had respiratory issues (rhinitis or asthma), and others had abdominal pain and/or vomiting. The eliciting dose of peanut ranged from 12.5 to 2000 mg, with a mean dose of 103 mg. As regards the five children with an anaphylactic reaction, their mean eliciting dose was 100 mg.

| BAT evaluation
BAT data were obtained for 39 OFC-positive patients and 48 OFCnegative patients because of nonresponder basophils (n = 2) and technical issues for the remaining two patients (Figure 1).

| sIgE evaluation
Whole peanut-sIgE and Ara h 2-sIgE levels were determined for 78 and 77 patients, respectively ( Table 1) Table 1). Receiver operating characteristic (ROC) curves for whole peanut-sIgE and Ara h 2-sIgE were also compared by means of the area under these curves (AUC, Figure 2). The AUC for Ara h 2-sIgE (0.88) was significantly higher than the one for whole peanut-sIgE (0.80, p = .0164). The ROC curve for Ara h 2-sIgE also showed that sensitivity is 47% for a specificity of 100%. Choosing then an integer TA B L E 1 Demographics of patient population. cut-off such that specificity is still 100% leads to 7 kU A /L for Ara h 2-sIgE. This means that all patients who did not have an allergic reaction during OFC had Ara h 2-sIgE ≤7 kU A /L. Hence, this cut-off value yields a specificity of 100% (95% CI: 91%-100%) with a sensitivity of 45% (95% CI: 29%-62%) and appears therefore useful to predict the need for an OFC.
Ara h 8-sIgE levels were determined for 56 patients, and no evidence of an association between Ara h 8-sIgE and PA was found (Table 1).

| Diagnostic strategy
All patients with an Ara h 2-sIgE >7 kU A /L were allergic to peanut ( Figure 3). This suggests that a value of >7 kU A /L could be used to establish a diagnosis of PA at this stage. However, this cut-off is not relevant in all patients who developed an allergic reaction during OFC since its estimated sensitivity is 45% (95% CI: 29%-62%; Figure 2).

| DISCUSS ION
In this study, we report the relevance of a model of composite biological markers to predict PA in children in 'real-life' settings. Our

F I G U R E 3
Relations between basophil activation rate, Ara h 2-sIgE levels and OFC results. All patients with Ara h 2-sIgE > 7 kU A /L had an allergic reaction during OFC. Among the remaining 22 patients with an allergic reaction during OFC, all but one had a BAT >6%. The one patient with a BAT ≤6% did not need to avoid peanut consumption because of an oral syndrome due to PR-10 manifestations. BAT, basophil activation test; OFC, oral food challenge.
allergic population was comparable to that of other studies. 2,15 The performance of BAT has been investigated in multiple food allergies, notably for PA. 9,16-18 Moreover, BAT has been suggested to be more accurate than peanut-sIgE in several studies on food allergy 9,10,12,14,18 and of particular interest in cases of multinut sensitisation. 11 Today, many peanut-sensitised patients still need to undergo an OFC in order to discriminate between peanut allergy and peanut tolerance. Thus, BAT has been investigated as an alternative to OFC. One study reported that BAT was able to reduce the use of OFC for PA diagnosis by two-thirds and proved particularly useful in cases in which specialists could not accurately diagnose PA with SPT and sIgE. 9 In our population study, peanut BAT using a previously published threshold of 6% had a sensitivity of 95% with a 95% CI ranging from 83% to 99%, that is, comparable with other studies ranging from 83% to 92%. However, its specificity of 54%, with a 95% CI ranging from 39% to 69%, was lower than that reported in the literature, where specificity ranged from 77% to 100%. We also tested the manufacturer's threshold of 15%, which had a lower sensitivity (85%, 95% CI: 69%-94%) for only mildly improved specificity (65%, 95% CI: 49%-78%). Therefore, we chose to use the 6% cut-off for our diagnostic strategy. However, these results suggest that BAT alone may lack specificity for paediatric patients and does not appear to be sufficient on its own to discriminate accurately between allergy and tolerance in peanut-sensitised children.
Therefore, we suggest that a staged diagnostic strategy, combining BAT results and sIgE levels, could overcome this lack of specificity ( Figure 5). Such staged diagnostic strategies have better performances than those based on BAT alone and, in some studies, better predict the severity and eliciting dose of an allergic reaction during an OFC. 10,19 As expected, no evidence of an association between Ara h 8-sIgE and PA was found, as Ara h 8 belongs to the pathogenesis-related protein (PR-10) family and is known to be associated with mild peanut-induced reactions or no reaction at all. 20 As regards Ara h 2, Ara h 2-sIgE considered alone is known to be very useful to predict clinical reaction during OFC. 21  writing -review and editing; writing -original draft.

ACK N OWLED G M ENTS
The authors are grateful to Nikki Sabourin-Gibbs, CHU Rouen, for her help in editing the manuscript.

FU N D I N G I N FO R M ATI O N
This study was supported by a grant from Rouen University Hospital (AOJC 2012 no. 2012-010).

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors report no potential conflict of interest.

PE E R R E V I E W
The peer review history for this article is available at https://