Evaluation of oral immunotherapy in hazelnut allergy: Pediatric experience in Toulouse

Abstract

In Europe, hazelnut allergy is the most common nut allergy with prevalence of 1.9%.¹ The current treatment remains total avoidance of hazelnut.^{2, 3} Oral immunotherapy (OIT) is promising but unvalidated for hazelnut, only two studies on this subject have been published.^{4, 5} We implemented a hazelnut OIT protocol in Toulouse University Hospital (France). We report our experience on 88 patients undergoing hazelnut OIT. Our objective was to evaluate the proportion of desensitised patients, indicated by a negative oral food challenge (OFC).

We performed a retrospective collection of patients who underwent OFC to hazelnut between 01/01/2016 and 31/12/2021 at the Children's Hospital of Toulouse. We included hazelnut-allergic children who either had a positive hazelnut OFC or a negative low-dose hazelnut OFC, associated with positive specific IgE (hazelnut, thermostable components nCora9 and/or rCora14) and/or a positive hazelnut skin test. After the 1st OFC, we implemented OIT; consisting of daily intake of hazelnut-containing products with progressive increase. We aimed to reach half the reactogenic dose as a maintenance dose until the second OFC (1490 mg hazelnut protein), on average 12 months after the first OFC.

Eighty-eight patients underwent hazelnut OIT, 69.3% were male and median age [interquartile range (IQR), range] was 8.1 years [(5.6–10.8), 16] at the time of 1st OFC. The median levels [(IQR), range] of IgEs were: hazelnut 6.7 kUA/L [(1.7–16.3), 99.5], nCora9 2.2 kUA/L [(0.6–10.9), 55.9] and rCora14 3.2 kUA/L [(1.0–13.5), 99.9]. The median reactogenic dose (IQR) was 289 mg [(102.2–640.2), 1470.67] of hazelnut protein.

Of the 88 patients included, 60 were able to perform 2 OFCs during the follow-up period. In the intention-to-treat analysis, 52.2% of patients were considered desensitised (n = 46/88). Of the children not desensitised (n = 42): 14 patients had a positive 2nd OFC and 28 children did not have a 2nd OFC (lost to follow-up (n = 17); discontinuation of OIT (n = 11)). In per protocol analysis, 76.7% of patients were desensitised (n = 46/60). The cumulative dose of hazelnut protein triggering the allergic reaction (IQR) was significantly higher in 2nd OFC 1490 mg (1490–1490) than in 1st OFC 187 mg (102–640) (p < .0001). For non-desensitised patients (n = 14), the reactogenic dose, in mg hazelnut protein, in 2nd OFC [810.2 mg (300.2–1490.2)] was significantly higher (6-fold) than in the 1st OFC [130.2 mg (102.2–289)] (p = .0143) (Table 1).

We collected the existence of adverse events during home OIT for 79 patients (Table 2).

The median (IQR) level of hazelnut IgEs f17, obtained at 1st OFC, in desensitised children [6.5 kUA/l (1.7–11)] was significantly lower than in non-desensitised children [21.8 kUA/l (3.5–58.9)] (p = .0174). Similarly, the IgE levels nCora9 and rCora14, at the 1st OFC were significantly lower in the desensitised group (p = .0265 and p = .0168 respectively).

Prediction thresholds were calculated for the probability of desensitisation for the variable IgEs against hazelnut f17, nCora9, rCora14. This predictive threshold is 16.3 kUA/l for hazelnut IgE with a positive predictive value (PPV) of 67% and a negative predictive value (NPV) of 85% (p = .007). For rCora14, below a threshold of 4.13 kUA/L, the PPV is 45% and above this threshold the NPV is 85% (p = .045). No analysis was performed regarding IgG4/IgE ratio.

In our study, 52.2% of the children were desensitised after a 1-year hazelnut OIT protocol. Most adverse events were mild, contrasting with the high discontinuation rate. Children with lower IgEs (hazelnut, nCora9 and rCora14) were significantly more likely to be desensitised (p = .0174, p = .0265 and p = .0168 respectively). We were able to establish thresholds for hazelnut IgE f17 at 16.3 kUA/L and for rCora14 at 4.1 kUA/L, above which the risks of OIT failure are high (NPV 85%). Although OIT is a long and restrictive treatment, our study shows effective desensitisation in a majority of children. New studies with a larger and more homogeneous number of patients would be desirable to confirm our results.

NC designed the work, made acquisition, analysis, interpretation of data and drafted the work. VG designed the work, made analysis, interpretation of data and substantively revised the work. ABC made interpretation of data and substantively revised the work. RP made acquisition and substantively revised the work. JK drafted the work. AMB substantively revised the work. LG substantively revised the work. MM designed the work, made analysis, interpretation of data and substantively revised the work.

The authors declare that they have no conflicts of interest.