Discussion:
A good understanding of the different trajectories of PA over time is important to safely diagnose and manage PA. The prevalence of PA at both the end of EAT and 7-11y was relatively stable with 2.1% at 7-11y and 1.9% at the end of EAT. There were two new cases of PA that developed after 36m and only 1 child outgrew PA by 7-11y. Children with persistent PA at 7-11y had significantly higher levels of SPT, peanut-sIgE, Ara h 2-sIgE and mast cell activation compared to children who were PS, with many biomarkers being diagnostic of PA by 36m.
At 12m of age the persistent PA children already had median SPT and Ara h 2-sIgE levels consistent with a PA diagnosis which only continued to increase over time. Studies have reported SPT of ≥6mm and Ara h 2-sIgE between 0.1-3kUA/L being predictors of persistent. (14, 15) Ara h 2-sIgE and Ara h 6-sIgE are the peanut components most indicative of true peanut allergy (16), which was consistent with our findings at the 7-11y time-point. The IgG4:IgE ratios were significantly lower in the PA group and specifically in the children with persistent PA at 36m and 7-11y. Overall, the MAT was suggestive of PA at the 36m and 7-11y time points in the children who had persistent PA at 7-11y. There were only two persistent PA patients who had plasma available from their 3m EAT study visit who had MAT performed (Table 2). Their median CD63 activation was 10.7% which is suggestive of PA at such an early age. The higher MAT at these time points reflected the higher levels of peanut-sIgE levels, which we know from previous work induces greater mast cell activation (7). These changes in biomarkers demonstrate that biomarkers that are high early in childhood and increase over time are indicative of persistent PA.
There were only 2 patients who developed new peanut allergy. Their SPT and peanut-sIgE biomarkers were initially low until 36m but increased over time so that by 7-11y they were consistent with a PA diagnosis. They were consuming peanut in early childhood but by 7-11y had stopped all peanut consumption which supports previous evidence that shows that patients who do not consume peanut regularly after a negative OFC are at higher risk of recurrent PA, which may have contributed to why these patients developed new PA. (17) Interestingly, MAT remained negative across time in these two children even at the 7-11y time point which differed from the children with persistent PA who had higher MAT at 36m and 7-11y. A possible explanation for this lower MAT in new PA is the quality of the IgE. Hemmings et al showed that IgE functional characteristics modify mast cell activation with higher mast cell activation resulting from higher peanut-sIgE levels, higher specific activity, higher diversity and higher avidity of IgE for peanut (18). It is possible that for those who had new PA later in childhood, the allergic immune response was not fully developed and sIgE had lower levels, specific activity, diversity and avidity for peanut allergens.
There was only 1 child that outgrew their PA by 7-11y confirmed by negative OFC. Although we cannot infer conclusions on trends in biomarkers overtime based on their results alone, the patterns observed were still interesting. This child had SPT and Ara h 2-sIgE suggestive of PA at 12m of age; although the SPT remained high at 36m, Ara h 2-sIgE level was negative. Their peanut-sIgE level was lower at all time points with a peak of 1.3kUA/L at 12m and then was negative by 36m and remained so until 7-11y. The IgG4:IgE ratio was also very high at the 36m and 7-11y time points which would be consistent with tolerance to peanut as seen in previous studies (12).
This study is unique in that it looks at the changes in PA in a population-based cohort of children over the span of a decade. The longitudinal nature of this study and the availability of biomarkers at the different time points helps to explain how PA is largely stable in later childhood. Our data demonstrates that high biomarkers in early childhood are associated with PA persistence which is consistent with previous findings (15). MAT has high specificity in identifying children who will clinically react to peanut (19) but this is the first study looking at MAT over time and for those with persistent PA, levels were raised by 12m. The utility of MAT was limited in children with very low levels of peanut-sIgE, like those who developed or resolved their PA.
The major limitation of this study is the small number of children in the sub-group analysis. Only 2 children developed new PA and 1 child outgrew PA which makes it difficult to draw conclusions. We had hoped to compare biomarkers predicting resolution of PA with persistence of PA but this was not possible in this cohort. There was also missing biomarker data in terms of baseline SPT (i.e. these were not performed for children randomised to the standard introduction group) and Ara h component-sIgE data (i.e. was only performed if peanut sIgE>0.1kUA/L). We were able to impute the component data based on sIgE levels but there were still some children who did not have data available. Also, as the children were all recruited from the EAT-On Study, definitions for allergic status and tolerance were based on the study protocol to allow for consistency in the data analysis. In an ideal setting, all children selected for the biomarker work would have had OFC to confirm their PA status at 7-11y of age.
To conclude, the rate of PA in this cohort of children was 2.1% at 7-11y. Children with PA at 36m and 7-11y have significantly higher SPT, peanut-sIgE, Ara h2-sIgE and MAT compared to PS children. These biomarkers are already raised between 12m and 36m of age. For those that develop new PA or outgrow their PA, the timing at which this happens likely occurred between 36m and 7-11 years of age but small numbers and low biomarkers prevented additional conclusions.