View Report Details

The prevalence and natural history of peanut allergy and investigation into its genetic, environmental and immunological determinants

Project Code: T07001

30/09/2003

Imperial College London
Lack, G

The main objectives of this project were to define the risk factors for the development of peanut allergy, the burden of paediatric food allergy in 7-8 year old children in the UK, risk factors for sensitisation to food at age 7-8 years and immunological differences that underlie T cell function in children with peanut allergy as compared to tolerant children. Given the great difficulties in answering these questions using selected case series of allergic children in a tertiary specialist clinic we utilised a birth cohort study. This large cohort study with its very high level of follow up and adherence provided the power to ask questions about prevalence and risk factors, and allowed investigators to carefully define the allergic phenotype and that of normal control children. The fact that most of the information was prospectively collected overcame the problems of recall bias that is present in many retrospective studies.

Important risk factors were identified for the development of peanut allergy. Thus, 35% of children with challenge proven peanut allergy had been exposed to soya formula in infancy (OR=3.15). The majority of children with peanut allergy had rashes in their joints and creases in infancy (OR=3.88) and the association with an oozing, crusted rash was even stronger (OR=24.62). The majority of children with peanut allergy had been exposed to several products containing arachis oil in infancy (OR=8.34). No less important were the negative findings in this study. Recall of maternal peanut consumption during pregnancy and breastfeeding was no different in the peanut allergic children’s mothers and mothers of non-allergic children. Information derived prospectively on breastfeeding and exclusive breastfeeding did not show differences between the peanut allergic and non-allergic populations. Importantly there was no evidence of specific IgE to peanut in any of the core blood samples of the children who went on to develop peanut allergy. Thus we hypothesise that exposure to low doses of peanut antigen through inflamed skin leads to allergic sensitisation. It may be that infant soya exposure (possibly also through the skin) may lead to cross-sensitisation as both peanut and soya belong to the legume family. To some extent the possibility of cross-sensitisation is supported by the strong association between sensitisation to soya and peanut. While all children with challenge proven peanut allergy in our study were tolerant to soya in their diet, 57% of these children were sensitised to soya based on demonstrable specific IgE in the serum. At five years of age, analysis of the blood samples in the ‘children in focus’ revealed that all subjects with specific IgE to soya had specific IgE detectable to peanut. This is an extremely high level of association not witnessed to the same extent between other allergens in related groups. One possible explanation is that crosssensitisation to peanut occurs through common T cell epitopes shared between soya and peanut.

A multi-variant logistic regression model was also used to examine multiple factors that may be associated with sensitisation to foods at the age of 7-8 years. Two very surprising findings emerged. Strikingly parental ethnicity was a strong risk factor for sensitisation to foods at age 7-8 years. Thus one non-Caucasian parent showed a strong association (OR=2.65) and two non-Caucasian parents had an even stronger association (OR=8.62). These findings are surprising since food allergies have been reported as being less of a problem in non-westernised countries where there is a higher non-Caucasian population. Clearly there is the possibility of confounding variables not accounted for in our analysis. Thus the non-Caucasian ethnic group could reflect perhaps an immigrant population that could show different social class demographics. However social class was not shown to have any association with sensitisation to food allergens. Another possible explanation is that children from non- Caucasian ethnic backgrounds in the UK are likely to be exposed to different dietary practices with differences in weaning practices. Important clues about the pathogenesis of food allergies could potentially be derived in the future from comparison of well-defined food allergies in different ethnic populations in the UK.

Importantly we found that passive smoke exposure was negatively associated with the presence of sensitisation in 7-8 year old children (OR=0.41 after regression analysis). This highly significant negative association was not explained by confounding variables such as social class. Nevertheless the question remains whether other confounding variables could explain this negative association. It is too early to conclude that smoking prevents the development of food allergies in childhood. However, interestingly passive smoking is associated with transient wheeze of infancy but not with persistent allergic asthma in childhood. There is the possibility that certain compounds present in smoke (such as nicotine) could potentially influence inflammatory reactions in the airways and skin and could also potentially affect T cell function. Further work is needed to characterise the association between passive smoke exposure and food allergy in childhood perhaps using a case controlled approach with careful delineation of food allergies in food allergic children and nonallergic control groups.

We determined that at the age of 7-8 years approximately 2% of children suffer from peanut and/or tree nut allergies. Most children with milk and egg allergy had outgrown this in the first 5 years of life and by the age of 7 or 8 there was very little food allergy to other foods. We know that we did not miss important food allergens because skin prick testing was done more on almost half of the ALSPAC population to a variety of food allergens including peanuts, nuts, egg, milk, fish, soya. Thus there was no appreciable level of fish or sesame seed allergy in the ALSPAC population at age 7-8 years. This is something that may change with time and one may expect to see more allergies emerging to fish, shellfish, sesame, fruits and vegetables during the teenage years. Given that most children of 7 or 8 will retain their peanut or nut allergies there is a possibility therefore that the food allergic burden to the population will increase substantially.

Importantly we showed that 40% of children with peanut allergy diagnosed at the age of 5 react to other tree nuts. This has never been proven before in the literature on the basis of food challenges. Tree nut allergy in association with peanut allergy has only previously been substantiated on the basis of skin prick testing, specific IgE testing and/or clinical history. Thus with almost half of peanut allergic children showing reactivity to tree nuts at the age of 5, a strong rationale is provided for the argument that children who are allergic to an individual peanut or tree nut need to avoid all peanuts and tree nuts in their diet. Furthermore one third of children who presented with peanut allergy in the first 5 years of life outgrew their peanut allergy by the age of 5.

Importantly we were able to establish 95% positive predictive values for specific IgE and skin prick test to peanut and tree nuts that allows prediction of whether a child will react to peanut and/or other nuts. This result was validated in both children in ALSPAC, a cohort community based study, as well as in children attending our specialist clinic (a tertiary allergy setting). We therefore have established what we believe to be “universally” applicable values for predicting food allergy in children. In practice this means that diagnosing food allergy will be easier and this may obviate the need for costly, difficult and potentially dangerous food challenges in some 30- 40% of the children who present to our clinics.

We also found important differences between peanut allergic and non-allergic children with respect to their T cell responses to peanut. We were able to develop a novel methodology using CSFE staining to identify antigen specific or peanut specific T cells that proliferated in response to peanut and lost their fluorescence. By a series of cloning experiments we proved that the CSFE low cells were indeed peanut specific. When we combined this method with intracellular cytokine staining we were able to demonstrate that peanut specific CD4+ T cells from peanut allergic individuals showed marked TH2 skewing in contrast to TH1 skewing seen in peanut tolerant children. Interestingly, in children who once had peanut allergy and had subsequently shown resolution of their peanut allergy the T cell responses demonstrated a cytokine profile identical to children who had always tolerated peanut.

Using PBMCs from the same groups of children we were able to do T cell proliferation kinetic studies in response to peanut and other control antigens. We found consistently that PBMCs from peanut allergic individuals proliferated to peanut with a high level of intensity on day 5 whereas PBMCs from non-allergic individuals showed less proliferation, which peaked on day 7-9. We have some preliminary data to suggest that this may be due to facilitated antigen presentation in peanut allergy. This will be the subject of future investigations.

In summary this project has allowed us to identify novel risk factors for peanut allergy. Identification of these risk factors may lead to new prevention studies in high risk individuals to prevent the development of peanut allergy. Similarly we have defined antigen specific immunological characteristics of T cell function in peanut allergic and non-allergic children. In future immunomodulatory trials, we intend to use such assays as immunological end points in order to determine response to new immunomodulatory treatments.

Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: