The essential criterion for diagnosis of food allergy is a response to an elimination diet, and other diagnostic tests are secondary to this. If there is allergy to a single food, exclusion should induce relief of all symptoms, and restore normal growth. For secure diagnosis, a positive response to challenge with the food antigen is strongly supportive of the diagnosis of food allergy. This is not always practicable in routine practice, if diagnostic tests were positive at diagnosis, and many parents may refuse challenge if their child has improved dramatically. Insurance companies, however, may require evidence of positive food challenge for reimbursement.
In addition to initial challenge for first diagnosis, subsequent challenge may be performed when it is reasonably likely that tolerance has been regained.
This is usually after 2 years, but often later, especially in multiply allergic children. The European Society for Paediatric Gastroenterology, Hepatology and Nutrition73 made recommendations for challenge criteria for children with predominant gastrointestinal symptoms, suggesting that blinded challenge is not always necessary in younger children, but that small-bowel biopsy should be performed in cases with failure to thrive or diarrhea. Skin prick testing (see later) may provide more specific guidance for children with immediate reactions.
In children with multiple food allergies, the response to elimination of single antigens is incomplete, and lengthy assessment with a very restricted diet is often required. Such situations may become complicated, particularly since the advent of Internet sites that implicate food allergies in an unfeasibly broad spectrum of disorders, and it may be difficult to persuade some parents to broaden their child's diet. For the child who manifests the symptoms of only non-IgE-mediated allergy, it may become extraordinarily difficult to obtain a clear picture of the true state of current true allergy. This situation is complicated by the minor immunodeficiency often associated with multiple food allergies, and the exacerbating effects of intercurrent viral illnesses.14,19,20 It is important not to perform food challenges while a child is systemically unwell. For cases of multiple sensitization, with delayed responses, it becomes logistically difficult to perform lengthy blinded challenges. Close teamwork with an experienced dietitian is essential in management of any such complicated cases.
The open food challenge is the most common form of diagnostic challenge, and is performed either in the out-patient clinic or in a day ward, depending on the severity and type of reaction expected.74 Many centers will use a graded challenge in those children who manifest immediate reactions, initially placing a single drop of milk or other antigen on the skin, then on the lips, and then giving increasing amounts by mouth, leaving a period of several minutes between each stage. For the child with a history of severe reaction, it is mandatory that there be adequate medical supervision with the presence of appropriate resuscitation and drugs. For those with a history of anaphy laxis, an intravenous line is usually inserted prior to the procedure.
For the child who makes delayed reactions only, the test may have to be completed at home for logistic reasons. False-negative results have been reported if late reactions are not taken into account. In a study of 370 challenges in 242 children, five exhibited mild immediate reactions that manifested only on the second day at home.75 These reactions were subsequently confirmed by skin prick test and double-blind placebo-controlled food challenge, and the authors suggested supervised feeding of antigen on the second day to identify late reactors. Hill et al17 identified very late reactions (up to 3 days) which required up to 120 ml to elicit. Recent data suggest that even later reactions may occur during cow's milk challenges, with the onset of gastrointestinal, respiratory or cutaneous responses beyond 1 week after reintroduction.76 Therefore, caution and judgement are needed to interpret food challenge tests, to prevent the missing of true late phenomena, but also to exclude over-reporting of symptoms by parents convinced that food allergy is responsible for all their child's various symptoms.
The 'gold-standard' challenge test, the doubleblind placebo-controlled food challenge is a cumbersome and time-consuming intervention, but is clearly necessary for assessment of cases with genuine uncertainty about sensitiza-tion.24,73,77 In cases of true uncertainty, the test can firmly establish tolerance or persistent reactivity in a way that no other test can. Its value is much lower when there are large numbers of foods to be tested, and the child makes delayed responses only. Normal day-to-day variation may then be overinterpreted, the situation can become stressed if performed on an in-patient basis, or intercurrent viral illness induces symptoms interpreted as due to the food allergy. This test requires extremely good teamwork between pediatrician, dietitian, pediatric nursing staff and the child's family. No universally accepted protocol exists,78 but in general the suspected allergen or placebo is given in disguised form, either by capsule or hidden in a liquid over a period of around 2 h. The child should not be receiving antihistamines and should not have received the antigen for at least 2 weeks, and should be well. A total dose of around 10g is normally taken as sufficient, although this would not pick up some of the delayed reactors identified by Hill et al,17 Carroccio et al76 or Caffarelli and Petroccione.75 Nevertheless, for immediate allergies, the test is highly specific and of very high prognostic value, with a suggested false-positive rate below 1% and false-negative rate below 5%.78
The skin prick test remains a cornerstone of the diagnosis of immediate allergic reactions, but is usually negative in those who manifest delayed reactions only.14,74 This is performed by placing a drop of the potential antigen on the skin (usually a commercially obtained solution, although some units use freshly prepared foods) and introducing it into the skin through a puncture device such as a lancet. A positive control (histamine) and negative control (saline) are used at the same time, and the results read at 15-20min. The test is based on the induced degranulation of cutaneous mast cells by antigen binding and cross-linking of their surface IgE molecules. The size of wheal elicited determines whether the test is regarded as positive. Usually a wheal of 3 mm or more is taken as a positive reading, although in young infants the test can be falsely negative, as the histamine-induced positive control is much smaller in infancy than later childhood, owing to the smaller numbers of cutaneous mast cells.79 Recent data suggest that the size of wheal obtained may be an important predictor of reaction, and large wheals (>8mm) may be so highly predictive of sensitization that unnecessary food challenges can be avoided.80,81 The study by Sporik et al,80 of 467 children aged from 1 month to 16 years, correlated the size of the wheal elicited to cow's milk, egg or peanut with clinical outcome. Using the traditional cut-off value of 3 mm was not found to be helpful in predicting a response to challenge, and many false-positive and false-negative results were obtained. However, using a cut-off value of 8 mm for cow's milk and peanut and 7mm for egg, they found no children with a negative response on challenge. Their calculations suggest that they could avoid clinical testing of 33% of milk-allergic, 56% of egg-allergic and 68% of peanut-allergic patients. These are potentially very important results, but clearly do require local replication, as testing solutions and technique will vary from center to center, and a negative test will not rule out a delayed non-IgE-mediated response to an antigen.
Roberts and Lack79 have extrapolated from the data from Sporik et al to suggest that a more precise risk value can be obtained using the Fagan likelihood nomogram,82 in which the risk of true reaction based on history is used together with the result of the skin prick to determine the overall likelihood of true sensitization. They provide an example where a child with low pre-test likelihood of peanut sensitization (headache and vomiting 4h after a peanut butter sandwich) would have a 0.2% likelihood of true peanut allergy with a 3 mm wheal, rising to 5% with a 6mm wheal and > 99% only at 10 mm, in contrast to a child with a high pre-test likelihood (urticaria and wheeze on two occasions within minutes of accidental peanut exposure), who would have a 15% likelihood of true peanut allergy with a negative skin prick test, rising to 70% with a 3 mm reaction and 96% with a 6 mm reaction.
The use of the skin patch (atopy patch) test, in which the relevant antigen is maintained against the skin under a sealed patch for 48 h, has been suggested to identify cases of non-IgE-mediated delayed allergy. A positive test is signaled by the finding of erythema and induration at 72 h. Combination of patch testing with either skin prick testing or specific IgE testing has identified significant numbers of food-sensitized children who may have been negative on classic testing.83,84 Of children with delayed reactions to antigen, 89% were identified by patch test in one study.83 This test has not become widely used and remains under evaluation. Recent reports do, however suggest that patch testing may be clinically useful, particularly in the presence of eczema.85,86 There are some logistic difficulties, in that the test should be read only after 3 days, requiring a second clinic visit, and some children will not tolerate an occlusive dressing for so long. It is also potentially open to artifact due to lack of supervision, and our unit has had at least one case of factitious illness, in which irritants were introduced beneath the occlusive dressing by the mother. However, it appears that the test induces a specific T-cell response to dietary antigen,87 and thus a properly performed patch test may be helpful in confirming delayed food allergic responses in cases of diagnostic difficulty.
Testing of specific IgE production to individual foods can be helpful. Formerly known as RAST testing, the technique is now usually based on enzyme-linked immunosorbent assay (ELISA) or the Pharmacia Cap system. It provides results that are complementary to skin prick testing, but at greater cost and with more delay. However, there is evidence that the information provided can be clinically predictive. Using the Pharmacia Cap system, positive predictive values of 95% were found for egg at 6kUA/l, milk at 32kUA/l, peanut at 15kUA/l and fish at 20kUA/l.88 These tests are frequently reported on a semiquantitative scale of 0 (no specific IgE) to 6 (high titers), with increasing likelihood of clinical relevance above level 3. However, there is wide variance between centers, and it is important to perform local audit to determine the clinical relevance of reported specific IgE. There is an increased likelihood of long-lasting allergy in children with higher titers of specific IgE for milk, casein or P-lactoglobulin.89
Potentially important prognostic information may be allowed in future by study of the specific sequence of epitopes bound by food-specific IgE. Cooke and Sampson90 identified two forms of IgE binding to ovomucoid (the dominant egg allergen). Binding to linear peptide sequences within the ovomucoid molecule was associated with long-lasting allergy, whereas binding to discontinuous sequences, which are brought into apposition only by protein folding, was seen in those who outgrew their allergies. This may explain the phenomenon whereby some children will be tolerant of cooked egg, where the tertiary sequence has been disrupted in the cooking process, but still react to raw egg white. For cow's milk, too, IgE reaction to linear sequences in casein, rather than the whey proteins a-lactalbumin and P-lactoglobulin, appears to be predictive of long-lasting milk allergy. IgE binding sites in specific sequences of aS1-casein, aS2-casein and K-casein were identified in those whose allergy was not outgrown.91,92 It is possible that this may provide the basis for future testing of milk-allergic infants, to identify those for whom immunotherapy may be necessary to prevent lifelong sensitization.
In vitro tests for lymphocyte response remains so far a research tool, and simple proliferation assays have given quite unreliable results. However, analysis of cytokine production patterns, using single-cell techniques such as ELISPOT or flow cytometry, represents a promising approach for the future. Lymphocytes from food-allergic children produce a pattern of cytokines associated with Th2 responses (IL-4, IL-5, IL-13) on antigen challenge, whereas those from healthy controls may make a more Th1-dominated response, with higher interferon^ production.93 However, studies have reported quite contradictory results when based on derived T-cell lines or clones, which may not represent true in vivo responses.
A promising novel technique has recently been reported, in which peripheral blood mononuclear cells were tagged with a fluorescent molecule called carboxyfluorescein succinimidyl ester, before stimulation with peanut antigen.94 This allowed simple identification of peanut-responding T cells by flow cytometry, and confirmed that T-cell responses to peanut in children with active allergies are indeed skewed towards Th2 (producing IL-4, IL-5 and IL-13, but not interferon-y or TNF-a), in contrast to findings of a Th1-dominated response with high interferon-y and tumor necrosis factor (TNF)-a in healthy controls or those who had outgrown their allergies.94 The authors reported similar skewing in milk- and egg-allergic children compared to controls, suggesting that this is likely to be a common mechanism in IgE-mediated food allergies. These findings imply that it is the host immune response, rather than a fundamental property of individual food antigens, that determines whether a Th2-skewed response occurs and allergy is induced.
It should be noted that mucosal responses appear quite different from systemic responses, and there are as yet no validated tests that might help in the diagnosis of non-IgE-mediated food allergies (Figure 22.5). TNF-a production is stimulated within the mucosa by milk challenge of allergic patients,38,95 which would not be expected from study of circulating T cells,94 but may reflect its release from mucosal mast cells. Studies of mucosal T-cell responses agree in showing that Th1 responses are maintained in intraepithelial,
Figure 22.5 Mucosal IgE responses may be distinct from those in other sites. (a) IgE immunohistochemistry of the duodenal mucosa of a child with milk-induced dysmotility, but negative skin prick testing, a serum IgE below 5 klU/l and no circulating specific IgE. Two types of IgE-positive cell are seen: round cells with a large nucleus and strong surface staining (IgE plasma cells) and irregularly shaped cells with intense granular staining (mast cells). (b) Confirmation that these granular cells are mast cells, as double staining for mast cell tryptase (red) and IgE (green) gives a resultant yellow color. Photomicrographs courtesy of Dr Franco Torrente.
lamina propria and Peyer's patch lymphocytes of food-allergic children.23,31,96
Was this article helpful?