Food Allergy Prevalence, Diagnosis and Impact: Unexpected Findings

In this issue, we present three studies which make important contributions to our understanding of food allergy epidemiology, food allergy diagnostics and food allergy's impact on populations. There are very few prior studies of changes in food allergy prevalence over time using robust methodology. Perhaps the only studies worldwide which have undertaken repeated, population-based assessments of confirmed food allergy prevalence are the UK Isle of Wight studies. In this issue, Carina Venter and colleagues summarise the findings of two birth cohorts on the Isle of Wight. They used these to evaluate whether a change in food allergy prevalence can be seen between a birth cohort born in 1989/90 and one born 12 years later in 2001/2 [1]. The significance of these years is that during this time, many reports have documented a sharp increase in hospital attendance and admission for food anaphylaxis in young children. Contrary to the popular narrative of a food allergy epidemic, Venter et al. did not identify a detectable change in food allergy overall between the two populations (Figure 1). This finding mirrors those of the US National Health and Nutrition Examination Surveys and the Australian HealthNuts and Melbourne Atopy Cohort studies, where no change in sensitisation to foods using blood specific IgE (US) or skin prick testing (Australia) to foods could be detected. Those studies compared children born in the 1970/80s with those born in the 1990s (US) and children born in 1993/4 versus 2010/11 (Australia) [2, 3]. The Isle of Wight findings are also consistent with a stable rate of fatal food anaphylaxis in national registry studies [4]. Taken together, these studies question whether food allergy has been increasing in high-income countries in recent decades.

If we look at two more studies published in this issue, we can start to understand why professionals and members of the public are convinced there is a food allergy epidemic, without good objective evidence to support that. The first issue is that food allergy diagnosis is difficult, so this health condition is susceptible to overdiagnosis. This is shown in the study of Chong et al. from Singapore, which involved evaluating the diagnostic accuracy of commonly used tests for milk, egg, wheat and peanut allergy in children [5]. They report poor diagnostic accuracy, especially for milk and wheat diagnostics, with the known issue of a high false positive rate for all testing modalities and allergens. Most children with food allergy do not have a supervised oral food challenge [6]. So, by relying on diagnostic tests and clinical history, and increasing the number of diagnostic tests performed, we may have allowed overdiagnosis to increase, fuelling a false food allergy epidemic [7].

The second issue is the social response to food allergy. There has been a marked increase in concern about food allergy in recent years, perhaps partly fuelled by increasing awareness and diagnoses of food allergy and marketing of adrenaline autoinjector devices, which tend to emphasise severe outcomes [4]. This increased concern can potentially lead to heightened anxiety and excessive avoidance behaviour, exacerbating social and dietary limitations and adverse mental health outcomes. It is thus of relevance to try to understand the intersection between food allergy and mental health. In this issue, Karim et al. investigated the mental health impact of food allergy using a large registry of twins in Sweden [8]. The authors used the Swedish Twin Registry and analysed data from over 30,000 twins born between 1992 and 2010 and evaluated from age 9–18 years. Data from this registry can be linked to prescriptions, birth records and their parents' records, allowing detailed epidemiological analysis. Karim et al. classified food allergy into three mutually exclusive groups: ‘parent reported without formal diagnosis’, ‘parent reported with formal diagnosis’ and ‘parent reported with adrenaline prescription’. They found only weak evidence for a relationship between one of the categories and adverse mental health outcome—the category ‘parent reported with formal diagnosis’. Due to the twin design and linkage with parent records, they were able to evaluate whether this association is likely to be causal or mediated by shared genetic or familial factors. Karim et al. found that any associations between food allergy and mental health were likely to be non-causal (Figure 2). This is consistent with previous mendelian randomisation work published in this journal, which used the UK Biobank to identify a lack of causal relationship between other allergic diseases and mental health outcomes [9]. Thus, food allergy may not directly cause mental health problems, but our familial and social response to food allergy has the potential to impact on mental health.

Taken together, the findings presented in this month's issue of Clinical and Experimental Allergy challenge some popular narratives around food allergy: that there is a childhood food allergy epidemic, that a positive allergy test indicates the presence of food allergy, and that food allergy, rather than our social response to food allergy, has significant adverse impacts on young people's mental health.

R.J.B. wrote the article. M.H.S. reviewed and approved the article.

R.J.B. declares payment for editorial work from Wiley and the British Society for Allergy and Clinical Immunology, and payment for legal advice in cases of food anaphylaxis and nutrition health claims. M.H.S. declares no conflicts of interest.