Article Text


IgE mediated food allergy: when is food challenge needed?
  1. P W Ewan,
  2. A T Clark
  1. Allergy Dept, Addenbrooke’s Hospital, Cambridge University Hospitals NHS Foundation Trust, and University of Cambridge Clinical School, Cambridge, UK
  1. Correspondence to:
    Dr P W Ewan
    Allergy Department, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, UK;

Statistics from

Commentary on the paper by Roberts (see page 564)

Roberts’ title is apt, but perhaps it should read challenging times for food allergy (or even for paediatricians).1 We are faced with an epidemic of allergy. The prevalence is rising, and recent data show over 30% of the population and 40% of children are affected.2 Food allergy is a particular—and seemingly increasing—problem in children, yet there is a dearth of trained allergists in the UK.3,4

Children have always suffered from egg or cows’ milk allergy, and it is well known that these resolve in the majority, often by 5 years of age.5,6 But new IgE mediated food allergies are appearing. The rise in peanut allergy is well documented. This was rare until the early 1990s, the first “big” series being published in 1996.7 The prevalence in children rose threefold over 4–6 years from 1 in 2008 to 1 in 70,9 and allergy to any nut now occurs in 1 in 50 (Dr Gideon Lack, ALSPAC data, personal communication).

We are faced with other new food allergies: allergy to fruits and vegetables, as part of the oral allergy syndrome (OAS) is now the “new” epidemic. These patients have rhinoconjunctivitis due to pollen allergy (especially birch), and through cross reacting proteins also react to fruits and vegetables. Children are also presenting with allergy to kiwi, sesame, and a range of other foods, distinct from OAS, but at lower rates.10

Good management of any disease requires knowledge of the natural history. While this is well known for egg and milk allergy, it is not for the newer disorders, for example peanut, kiwi, and sesame allergy.


Diagnosis can usually be made from the history backed up by evidence of specific IgE antibody. It therefore seems surprising that challenge is often suggested to confirm food allergy.1 Roberts rightly emphasises the importance of history. This is key to diagnosis, and is easier to interpret in immediate onset food induced reactions. The time interval between ingestion and onset is short, symptoms are typical, and in severe reactions progress is usually rapid and respiratory symptoms dominate.


Skin prick tests are immensely helpful but interpretation is vital. A positive skin prick test supports the diagnosis suspected from the history. A negative refutes this. Of course there are always caveats; interpretation should be in light of the history as emphasised by Roberts1 and valid allergen extracts should be used. While commercial extracts are excellent for many food allergens, they are of no value for certain foods (for example, prick-prick tests should be used for fruits). Technique of skin testing, including attention to negative controls, is also important, and this is difficult to achieve unless done regularly.

An important distinction is the difference between sensitisation and clinical allergy. Sensitisation is the presence of specific IgE; this may occur without clinical expression of allergy but is a necessary precursor. A positive test shows the presence of specific IgE. The “false positive” Roberts refers to is not really a false positive—specific IgE is there. It is the consequence that varies: allergy or no allergy. Failure to understand this is why allergy testing is sometimes viewed with scepticism.

Using peanut allergy as an example, skin tests of 8 mm or greater, are almost always associated with allergy.11,12 If the skin test is in the 3–7 mm range, approximately half will be allergic and half tolerant, so this is a grey area where interpretation is important. Thus a skin test in this range with a typical history confirms allergy. Similarly cut-offs predictive of allergy have been determined for other foods: for egg this is 7 mm and for cows’ milk 8 mm.12 There are smaller cut-offs in children under 2 years.12,13

The general conclusion is that history is important and this is best supported by skin prick tests. Skin tests have excellent sensitivity of >95%, so that a negative result is very helpful, but a low specificity of 50% (if history is not taken into account). “Blind” testing, where there is no history, is unhelpful.


The reality is that paediatricians mainly have access to blood tests (serum specific IgE tests, commonly referred to as “RAST” or ImmunoCAP). However they are not as satisfactory as skin tests and do not give identical results. Negative “CAP/RASTs” can be falsely reassuring: for example, in nut allergy this was found in 19% with a convincing history and positive skin test.11 Roberts’ criticism of all specific IgE tests seems to be based on serum assays rather than skin tests. Similarly data for egg and cows’ milk show a few children have a “100% diagnostic” skin test level and negative serum specific IgE levels.12


Usually a child with a negative skin test does not need a challenge; those with strong positive skin tests have allergy and in those with intermediate positives the diagnosis needs to be informed by the history. The history is usually sufficient in the case of acute reactions. Graham Roberts and Gideon Lack pointed out the power of combining history (pre-test probability) with skin tests (likelihood ratio) to give a post-test probability of food allergy.14 There is no doubt that experience, particularly of large numbers of the same type of case, is important.

The approach is different with chronic symptoms. With a positive skin test and no history of immediate reactions to that food, for example, a child with eczema or asthma, the correct approach would be a trial of dietary exclusion then reintroduction, at home with a symptom diary. Another scenario is a child with a positive skin test who is eating the food without reaction. Neither of these requires a challenge.

The main clinical indication for challenge is to show resolution of allergy,15 for example a child with a clear diagnosis of nut allergy (history supported by positive skin test) who later, after nut avoidance was found to have a negative skin test. However, in practice it would be advisable first to repeat the skin test, after a period of time.

There is no consensus on indications for food challenge, but there is a trend to using this in obvious food allergy. It is time consuming, needs training, and there is an increased risk if patients are poorly selected. Roberts reported a 10.2% incidence of severe reactions during food challenges but selection criteria were not described.1 Morisset et al, in a study designed to identify threshold doses for reactivity, in patients with a convincing history and positive skin test, found serious clinical reactions including hypotension, for egg, milk, peanut, and sesame.16 Therefore it seems wise to avoid unnecessary challenges. It is of interest that there are no studies comparing an allergist’s diagnosis with challenge. The data showing a discrepancy between perceived allergy/intolerance and challenge, are based on questionnaire or subsets from questionnaires.17,18

Food challenges are often seen as an easy solution, but this is not the case. They are not straightforward and a number of variables need to be considered. These include factors pertinent to the allergen, for example, dose, matrix (what the food is mixed in19), whether cooked or raw; and to the patient, for example, dose20 and masking (which is often difficult to achieve21), current asthma control, and other active allergies. The starting dose needs to be tailored to the patient, taking account of the severity of the reported reaction(s) and the dose causing this.20,22 As Roberts points out, anxiety may produce subjective or even objective (erythema or vomiting) symptoms. There is no agreed protocol and no two centres do this in the same way. Do you do a double or single blind or open study? It should not be delegated to inexperienced staff, and needs to be done regularly—it is not suitable as a one-off procedure. Assuming patients are appropriately selected, where one is expecting a negative or at worst a mild reaction, an intravenous cannula is not usually required.


Even this is not straightforward. All one can say is that the child can tolerate that dose, in that form, on that day. The reality is that the majority of patients can then tolerate the food. Rarely will a patient react to open challenge with a food after a negative challenge with 8–10 g,22 the assumption being a larger dose was ingested.


The key to dealing with patients with food allergy is having a good clinical knowledge of allergy. This means having experience in allergy history taking, allergens, natural history, and interpretation of allergy tests (whether SPT or RAST) before contemplating food challenge. Food challenge is not for the inexperienced. This leads to bizarre practice with the risk of misdiagnosis. In reality the need for food challenge is reduced by good clinical practice. This is important as food challenges are time consuming and paediatricians are inundated with a heavy allergy case-load and effort should be focused to maximise efficiency.3

If the approach suggested above is followed, the need for food challenges is reduced. However, there is much to be discovered and understood and clinical research is needed. Clearly food challenge tests play an important role in research.

There is a huge need for more paediatric allergists and general paediatricians with an interest in allergy.3,4 The problem is broader than food allergy alone because many children have multi-system and often severe allergic disease (including asthma, eczema, and rhinitis).7,10,23,24 Proposals to improve allergy care have been suggested.3,4,24

Commentary on the paper by Roberts (see page 564)


View Abstract


  • Competing interests: none declared

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.

Linked Articles