Article Text


History, blood tests or skin prick testing? Is it possible to predict the severity of allergic reactions in children with IgE-mediated food allergy?
  1. Lotte Harleman1,
  2. Adrian Sie2
  1. 1 Royal Aberdeen Children's Hospital, Aberdeen, UK
  2. 2 Wishaw General Hospital, Wishaw, UK
  1. Correspondence to Dr Lotte Harleman, Royal Aberdeen Children's Hospital, ST5 Paediatrics, Westburn road, Aberdeen AB25 2ZG, UK; lotteharleman{at}, l.harleman{at}

Statistics from


A 5-year-old boy with a history of an itchy rash and lip swelling following peanut ingestion several weeks ago comes to see you in your outpatient clinic. His mother would like to know if her son will have a more severe reaction if he is exposed to nuts again. What advice do you give her and which, if any, investigations should be done?

Structured clinical question

In a child with a history of an IgE-mediated allergic reaction to food (patient), are (intervention) blood tests, skin prick tests (SPTs) or the medical history helpful in determining the severity of future reactions (outcome)?

Search strategy

A Medline search 1946–2014 was performed, as per table 1. The same search terms were used for the Cochrane library, with nil relevant articles, and also for an Embase search which found one further original research paper.1 Anaphylaxis was searched separately as (predic* adj3 anaphyl*).mp in Embase, Cochrane and Medline, finding one new review article2 and one new original research article.3 Articles were excluded if they did not relate to food allergy or children.

Table 1

Medline 1946–2014

The 16 original research articles included are summarised in table 2.

Table 2

Summary of articles

The findings of the relevant review articles2 ,4–6 and recommendations from the Royal College of Paediatrics and Child Health care pathway for food allergy7 and the National Institute for Health and Care Excellence (NICE) guidelines on food allergy in children and young people8 are discussed further in the Commentary section.


Immunoglobulin E (IgE)-mediated food allergy is common in children and is on the increase.8 There are studies that give a range of positive predictive values (PPVs) for SPT wheal size or specific IgE (sIgE) levels, aiming to help in diagnosis and reduce the need for food challenge in hospitals.23 The prediction of severity of allergic reactions is less studied and the NICE guidelines recommend further research in this area.7 ,8

Historically accepted risk factors for severe allergic reactions are asthma, older age and a history of a previous severe reaction. These risk factors were accepted on the basis of older population-based data and case series of fatal anaphylaxis.5 ,8 ,24

The questionnaire-based population studies18–21 included in this study also support asthma as a risk factor for severe reactions. However, three studies looking at children undergoing food challenge1 ,12 ,22 show no difference in reaction severity for a history of asthma or other atopy. One explanation for conflicting results may lie in the different circumstances of the reactions, particularly the likelihood of the ingested dose of allergen on accidental exposure in the community being higher than in a typical food challenge. One study comparing allergic reactions in the community with double blind placebo-controlled food challenge (DBPCFC) finds only a weak association between reaction severity on DBPCFC and in the community.25 Furthermore, population studies include children with previous anaphylaxis, whereas most food challenge research5 ,22 excludes these children for safety reasons. Another factor to consider is that asthma is a common condition, reported in 50% of children with food allergy by Spergel et al. 20 So although the questionnaire-based population studies and analysis of fatal reactions show that severe reactions tend to occur in children with asthma, as asthma is so prevalent in children with food allergies, this also means that most food allergic children with asthma will not necessarily develop anaphylaxis. It is also important to note that there is significant variation in the severity of asthma and most research does not differentiate this. Summers et al, however, differentiated this variation and classified atopic children into severe or mild, depending on whether their symptoms were under control. They found that severe asthma gave a significantly increased relative risk (RR) of 6.8 for life-threatening bronchospasm in comparison with patients with milder asthma symptoms who had an RR of 2.7 for a severe reaction.21 This underlines the importance of achieving good asthma control in all children, particularly those with a diagnosed food allergy.

In terms of the importance of the history of reaction predicting future severity, Wainstein et al 22 show no association between previous reactions and severity of the reaction during food challenge. This is the only study where the food challenge protocol allowed for further challenge doses despite onset of objective symptoms, as most other protocols stop at first onset of objective symptoms. There are limitations to this study such as lack of blinding and a small sample size. However, it is interesting to note that Wainstein et al 22 found that 21 of 27 children, or 78% of the children who completed the oral food challenge (OFC), developed anaphylaxis. This suggests that the dose of the allergen consumed is also important in determining the reaction severity. Other much larger studies12 ,15 ,19 ,20 show a similar severity of reaction on re-exposure. Calvani et al 11 found that the organ system affected in the initial allergy history correlated with the reaction type during acute anaphylaxis. Abdominal symptoms were associated with increased risk of collapse, a finding also noted by Hourihane et al. 14 Spergel et al 20 found that the overall most common reaction on OFC was cutaneous at 43%, with anaphylaxis counting for 14% of reactions. A total of 218 children had a history of initial cutaneous reaction and 56% of these children also had a further cutaneous reaction on OFC but 12% of children with an initial cutaneous reaction developed anaphylaxis on OFC, showing that a severe reaction is possible following any initial clinical presentation.

Most food challenge studies did not find that older age was significantly associated with reaction severity.10 ,22 Cianferoni et al 12 found that age above 5 years had a PPV over 70 for anaphylaxis. Hourihane et al 14 included adults and children and found more serious reactions in adults.

In 7 of the 11 studies reviewed above which looked at level of sIgE and/or SPT wheal size, a statistically significant correlation with severity of reaction was found. However, there is a wide range of results. For example, in Hourihane et al 14 the range of SPT results in children with severe reactions (4–15 mm) is entirely within the range of those with mild reactions (4–20 mm). Three studies8 ,19 ,20 found no association and Clark et al 13 found a significant association with SPT results in severe reactions to brazil nuts only. There is some limited quality data from both Michaud et al 3 (level 4) and Neumann-Sunshine et al 15 (level 3b) that highly sensitised children with sIgE >100 kU/L for peanut report more severe reactions. Neumann-Sunshine et al calculated an OR of 3.44 for anaphylaxis based on the reactions of 105 children with sIgE >100 kU/L. This suggests that children who are highly sensitised to peanuts are at higher risk for anaphylaxis. However, first, only a small proportion of children in the study are highly sensitised. Second, some children from this highly sensitised group do not go on to have severe reactions, whereas children with lower levels of sIgE do.15 Overall, despite associations of higher sIgE levels and higher SPT sizes with more severe reactions, the results are not of clinical significance for predicting future reaction severity in most children.

Component testing for sIgE to individual peanut proteins may be used in the diagnosis of food allergy to try to distinguish sensitisation from allergy with some success, although not necessarily with more accuracy than standard SPT.26 Astier et al hypothesised that there may be a correlation between the components to which an individual is sensitised and the reaction severity. They found that patients monosensitised to rAra h 2 on SPT have a lower disease severity score than patients who are polysensitised to rAra h 2 and rAra h 1 and/or 3 in a case–control study.10 This is the only paper looking at disease severity score and sensitisation on SPT for peanut components. Other research shows no correlation with level of sIgE for Ara h 2 and disease severity score.27 As the molecular understanding of allergy increases, differentiating between allergen structures may become helpful in determining the severity of a food allergy in the future, although there is currently not enough evidence to support the use of component testing in predicting the severity of allergy.

Median basal tryptase levels were found to be higher in children with anaphylaxis in a case–control study (level 4 evidence) by Sahiner et al. 17 However, there was significant overlap in levels between cases and controls and only four children had high tryptase levels, two of whom were non-allergic controls. There is currently not enough evidence that this test will help predict an individual's reaction severity.

In summary, although SPTs and blood tests may be useful in the diagnosis of food allergy, there are currently no tests that will accurately predict which children will go on to develop a severe allergic reaction.

Given the difficulty in predicting future reactions, it is important to educate families about food allergy. Some of the studies reviewed above rely on accidental exposure for their data on reaction severity and report high rates of postdiagnosis exposure, up to 50% over approximately a 5-year period.15 ,19 Avoidance of food allergens is clearly difficult for children and their families. Following accidental exposure, only low numbers of patients (34%) with severe reactions had adrenaline administered.15 Patient education is therefore the key clinical priority, as highlighted in a recent review on predicting reaction severity in peanut allergy.6

Clinical bottom line

  • We cannot accurately predict which children with food allergy will have severe reactions in the future.

  • The patient described above should have a thorough history taken and skin prick testing and specific immunoglobulin E testing, as per National Institute for Health and Care Excellence guidelines, to confirm the diagnosis.

  • Following diagnosis of a food allergy, it is important to educate the family concerning allergy avoidance and management of accidental exposure.


We would like to thank Paul Manson, clinical librarian at NHS Grampian for his technical assistance with the literature search.


View Abstract


  • Contributors LH wrote the article originally as part of an assignment to do a literature review for the University of Glasgow Certificate in Child Health. AS was the supervisor for this review and developed the clinical question. He also reviewed and helped modify drafts of the article.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; internally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.