Series Originator: Edward Kaminski Series Editor: Stephen Jolles
J. O'B. Hourihane, Paediatrics and Child Health, University College Cork, Cork, Ireland. E-mail: j.Hourihane@ucc.ie
Allergic conditions are common, with asthma being the most common chronic illness in childhood in most developed countries. Some 80% of asthmatic children are sensitized to aeroallergens, usually indoor animal dander and house dust mite. Some 80% of asthmatics also have rhinitis. Rhinitis and eczema receive less medical attention than asthma, but they can cause long-term morbidity and have substantial direct and indirect economic costs. Food allergy and anaphylaxis are increasingly recognised and are usually easily diagnosed and managed. Clinicians can use in vivo and in vitro measurements of allergen-specific immunoglobulin E to better time reintroduction of implicated foods. Specific parenteral and sublingual immunotherapy is widely practiced internationally but is uncommon in the UK. It may alter the natural history of aeroallergen reactive diseases in the upper and lower airways. Specific oral tolerance induction represents the current cutting edge in clinical allergy research. It remands resource intensive at present and cannot be adopted into routine clinical practice at this time.
Two clinicians working with children in Vienna in 1906, Clemens von Pirquet and Béla Schick, suggested the term allergy (from Greek allos = other and ergos = reaction) to describe an inappropriate, or out of place, reaction and an allergen as a protein that caused this ‘supersensitivity’. Allergy in childhood is common. The cumulative prevalence of asthma in childhood may be 39% . The indolent and chronic effects of allergy in childhood on school performance, socialization and physical activity are often overlooked when parents and teachers are more alarmed by the unlikely but unpredictable prospect of anaphylaxis because of food allergy.
It is recognized that many allergic children may see an immunologist with an interest in allergy sooner than they will see a paediatrician with the same interest. The unique additional aspects of health care for children that must be borne in mind are, briefly: the issues of nutritional adequacy, quality of life and occasionally overwhelming parental anxiety. With infants and children the individual needs and health-related behaviour of the affected child and the whole family must be considered to ensure adequate growth: not only nutritionally for somatic growth, but for social and emotional maturation [3,4]. This review will present an approach to a child with allergy as seen in routine paediatric allergy practice, including advice on allergen avoidance, prescription of rescue medications and immunotherapy.
The discovery of immunoglobulin E (IgE) by Ishizaka and Ishizaka  in the 1960s and the development of practical laboratory means of measuring the very small amounts of total IgE and allergen-specific IgE in blood by Wide and colleagues [6,7] are landmarks in the history of immunology. Today there are two common ways of demonstrating IgE antibodies used in everyday clinical care: in vivo allergen skin prick testing (SPT) and in vitro measurement of serum allergen-specific IgE, using the widely available ImmunoCAP® (Phadia, Uppsala, Sweden) or other emerging systems . Clinicians must be aware of which system their laboratory is using, as most of the international literature relates to Immunocap® and its predecessors and there are very few direct comparisons of each test system's performance characteristics .
Allergen skin testing
The SPT was first described by Blackley  in 1873 as a means of demonstrating pollen sensitization. This is a safe, practical and highly patient- and parent-acceptable way to look at allergen sensitivity in infants and children. A small amount of standardized allergen is introduced epicutaneously, using a standard single- or double-tined lancet. It is important to note it is no longer accepted practice to use a small-gauge hypodermic needle, or to use a single lancet for multiple allergens . Allergen cross-bridging of IgE fixed on mast cells results in release of vasoactive histamine and other mediators of inflammation. Within 10 min a palpable itchy ‘urticarial’ papule or wheal appears, not unlike a mosquito bite. The wheal is measured using a ruler, and recorded either as a mean of two perpendicular diameters (recorded in mm) or as an area, recorded in mm2, using a computer-linked laser reader. The older practice of comparing the wheal size to that of response to the histamine control is no longer supported, as clinicians can now use absolute wheal size expressed in mm to predict clinical reactivity, as assessed in a food challenge [12,13].
The SPT controls
However, it is still important to always use a histamine positive control, to ensure that the child is actually able to mount a wheal and flare response (thereby validating any wheal response that is elicited by an allergen), and is not taking an anti-histamine, which would block such a response. A negative (saline) control is always used to ensure that the child does not have dermographism or a pressure sensitivity.
The SPT can be undertaken in infants and young children and is noted for its safety and acceptability. It can be performed in low-risk children in home or school settings [14,15]. Most adverse reactions to SPT occur in subjects with unstable allergic conditions, particularly inhalant allergen-associated asthma. These children are identifiable by history and should have SPT performed in a hospital setting .
In older children and adults the volar (palmar) aspect of the forearm(s) is used; in infants the back is the best, as the infant can be held and comforted against a parent or assistant.
Allergen SPT is difficult if there is widespread eczema. It is said to be unreliable soon after an anaphylactic episode (when the mast cells have all degranulated and require ‘re-stocking’). Apart from anti-histamines, common drugs have little effect and oral steroids and topical calicineurin inhibitors for eczema such as tacrolimus appear to have no adverse effect.
Some allergens, notably those from fruits and vegetables, are unstable and the best way of demonstrating sensitivity to these [in such cases as the oral allergy syndrome (OAS)] is by using the fresh fruits and vegetables. These are then cut and either a drop of juice is placed on the skin and the lancette is pushed through this or the lancette is inserted into the fruit and then directly into the forearm (‘prick–prick method’). Because there is still no universally accepted standardized unit of reactivity for SPT solutions, some units never use commercially available SPT solutions and use only fresh foods, such as cow's milk, egg white, etc.
Scratch testing is not used in clinical practice. Intradermal testing, where a much larger dose of allergen (approximately 200 times greater than SPT) is injected into the skin, is now generally reserved for the identification of drug sensitivities, notably anaesthetic agents. End-point titrated SPT is not part of routine clinical practice but is making a comeback in the setting of deciding start doses for specific oral tolerance induction (discussed below).
It is important not to use too many allergens when testing as SPT is specific, particularly for common food allergens, but it is not sufficiently sensitive. The prevalence of an allergic disease in the population being tested has a strong effect with 50% sensitivity noted in population-based studies , but much higher levels up to 90% in referred populations. Testing should be restricted to those allergens that are implicated by history [e.g., in asthma test for house dust mite (HDM), cat, grass] or where there is a high likelihood of co-existing allergies (in infants test for milk and egg, e.g. or in older children it is completely justified to skin test peanut allergic children with tree nuts).
We and others have shown that above a certain wheal size for milk, egg and peanut (8, 7 and 8 mm respectively), a universal positive reaction is seen on open food challenge [19,20]. However, this did not hold for wheat, fish or soya. These published ‘decision points’ are used widely in practice to decide who would or would not pass a food challenge. They have been developed usually in highly selected populations of subjects who have usually reported clinical reactivity, and/or have undergone formal food challenges (see below); therefore, they are not easily applicable to subjects with high specific IgE levels (e.g. to egg or peanut) but no history of reactivity (a common reason for referral). Furthermore, they may not be relevant to less selected populations seen in routine practice. It is recommended that each centre or unit establishes its own decision points, using the allergens and tests they use routinely .
In vitro IgE
Our experience is that measuring total IgE rarely adds to the diagnostic process in an allergy clinic, as most children seen have raised levels (the prevalence of atopic disease in this referred population must be approaching 80–90%; see above discussion of SPT). Total IgE is not an allergy screen, but may be useful in the interpretation of specific IgE results when the total IgE is very high. It is important to provide accurate clinical details on the request form to allow the laboratory to offer optimal testing and advice. This is a prime area for repeated audit of practice and outcomes in clinical allergy and immunology services.
It is often a traumatic experience for children to have blood taken, and this should be respected when laboratories are dealing with paediatric samples. Many laboratories will save the serum for a number of weeks to allow subsequent testing to be carried out on the same sample when necessary.
The system used in most UK laboratories to measure allergen-specific IgE in blood is the ImmunoCAP® system. This is a modification of the original radioallergosorbent test (RAST). There is a wide choice of commercially available native allergens and individual ‘component’ allergenic proteins and recombinant proteins. The results are expressed in international units of IgE (kUA/L) and no longer in RAST classes. There have been studies linking the risk of clinical food sensitivity to threshold values of allergen-specific IgE , and the same caveats as mentioned above for SPT apply to specific IgE levels.
Specific allergic conditions
The largest proportion of the infants and children referred for evaluation are for the clarification of a suspected allergic adverse reaction to food. The prevalence of food allergy has been measured in 6-year-old children on the Isle of Wight, and while the reported prevalence was 11·8%, with increasingly stringent testing (allergen SPT, open challenge, double-blind challenge), the prevalence fell to 3·6%, 2·5% and 1·6% respectively . The story given is usually clear-cut. Exposure to a food (e.g. milk) to which the child is sensitized results in an immediate (within minutes) reaction causing a constellation of symptoms [lip tingling; lip swelling; lip, facial, body urticaria; redness; tiredness; vomiting; laryngeal swelling, stridor; dry cough, wheeze, severe wheeze, cyanosis; older children report a sensation of impending doom; collapse and hypotension (although cardiovascular symptoms are unusual in children)]. The ‘usual suspects’ in young infants and children are cow's milk, egg and peanut . Occasionally, reactivity to soya and wheat can be verified, but our personal experience is that this is uncommon. Most IgE to wheat in older children reflects cross-reactivity with grass (wheat is a grass). In older children reactivity to tree nuts, fish and shellfish become more common. The cited foods, along with lupin, mustard, molluscs, sulphites and seeds, form the list of ingredients that must be declared on ingredient labels in the European Union, irrespective of amount or concentration (see Table 1).
Table 1. List of allergens that must be declared on food labels in the European Union, irrespective of amount or concentration.
European Commission directive 2006/142/EC, amending Annex IIIa of Directive 2000/13/EC.
However, there is a myriad of foods (and also cross-reactions, e.g. between banana, kiwi and latex; cow's milk and goat's milk) which can cause reactions, so the clinical history needs to be taken carefully. Reactions caused by less common allergens such as potato, peach, grape, etc. should not be overlooked. Discussion of non-IgE-mediated reactions to foods are beyond the scope of this paper and interested readers are referred elsewhere [23,24].
There are important regional variations which reflect local diet. Lentils and peach are common allergens in Spain but not in the United Kingdom (Europrevall, unpublished data). OAS symptoms are common to hazelnuts in northern Europe, where cross-reactive allergens from birch trees are relevant, but hazelnut anaphylaxis is more common in southern Europe, where birch is absent and a lipid transfer protein is the dominant hazelnut allergen .
It is common for parents to ask if it is true that subsequent reactions are more severe than the presenting one. The Cambridge group has shown that subsequent reactions to peanut and tree nuts tend to be similar. When under expert care, nearly all patients self-treat appropriately when they next experience an allergic reaction [26–28]. Variability of reactions as has been shown for peanut, for example , more probably reflects the interplay of co-factors such as dose, viral infection, the new onset of asthma after the primary diagnosis of food allergy, the effect of exercise  and, in adults, use of alcohol and prescribed medication .
Physical examination is not usually helpful in the diagnosis of food allergy, but the presence of eczema, Dennie-Morgan folds (multiple wrinkles beneath the lower eyelid) and an allergic crease (see Fig. 1) are suggestive. In older children it is critical that asthma is properly identified and controlled.
Only a third of reactions reported in infants can be substantiated by challenge , as other mechanisms of adverse reactions may be operating. Therefore, in some health systems (e.g. Finland) a definitive diagnosis must be made using a food challenge if the costs of dietary substitutes such as extensively hydrolysed formulae are to be reimbursed by State health insurance (they can be expensive). Unsubstantiated exclusion diets may be nutritionally inadequate [33,34]. In the United Kingdom early food challenge is indicated if there is clinical doubt about the diagnosis or the test results conflict with the history. Food challenges are also used to investigate asymptomatic sensitization to associated foods detected at diagnosis, and especially to determine resolution of food allergy. It requires experience to time a challenge optimally so that exclusion diets are not sustained for too long. Food challenges are easily blinded, but this is rarely necessary in routine paediatric practice, in contrast to research protocols where the gold standard test of double-blind, placebo-controlled challenge is required. These double-blind tests have a bad public image, but are not usually more labour-intensive except for requiring 2 half-days of day ward attendance instead of 1. Agreed protocols have been established .
Once food allergy is confirmed, avoidance of that food is necessary. This can be quite easy if the food is distinctive and easily identifiable (e.g. kiwi) but can be impossible, without expert help, if it is common and difficult to detect (milk, egg, peanut). Experienced dietetic input is critical, especially for infants and smaller children to ensure nutritional adequacy and in older children with multiple food allergies. Soya milk is neither advised for treatment of young infants with cow's milk allergy nor for prophylaxis in asymptomatic infants considered at high risk of atopic disorders [36,37]. An extensively hydrolysed milk formula or an amino acid-based formula which have reduced and no allergenic activity, respectively, are required. Partially hydrolysed formulae are often used in Europe. There is no role for other mammalian milks, such as goat's milk, in the treatment of cow's milk allergy .
Natural history of food allergy
At a population level most infants allergic to milk and egg become tolerant to these foods by school age, although resolution rates in referred populations are lower . Peanut allergy resolves in about 25% of cases [39,40]. Longitudinal follow-up is therefore needed to distinguish ‘resolvers’ from ‘persisters’. Sequential SPT and specific IgE levels may show a reduction in sensitivity that heralds the acquisition of tolerance , although the underlying immunological mechanism has not yet been determined. The area of specific oral tolerance induction is very active at present , but there is as yet no consensus regarding protocols and eligibility criteria.
Eczema and food allergy are common bedfellows in infancy and there is a strong association between severe eczema and food allergy in children under 1 year of age . This association attenuates over time. It does not appear useful to screen older children with eczema for food allergies. Parents of these older children may despair for resolution of eczema or may be worried about the toxic effects of topical treatments, but food allergy is rarely the cause . In infants aggressive treatment should be aimed at all possible aggravating factors (using extensively hydrolysed or amino acid formulae in infancy; potent topical steroids or calcineurin inhibitors, antibiotics, emollients and moisturizers, anti-histamines, to decrease itch) to improve the skin quickly; once ‘healed’ it is less prone to flare. The diet can then be liberalized one food at a time and the response of the eczema to foods noted. In older children HDM allergy plays an important role, although the distribution of eczema may be different (exposed hands, feet and neck). Pollen exposure can also exacerbate eczema and increased eczema medication use (as well as rhinitis and asthma treatment, see below) may be needed during the pollen season. Preventive HDM reduction measures are not effective on a population basis but can be unpredictably effective for individuals. Immunotherapy for HDM-sensitized subjects with eczema has a moderate additional impact [45,46].
Allergic conjunctivitis and allergic rhinitis are common and have significant socio-economic effects because of poor school performance, decreased quality of life and other direct and indirect health-care costs . The major allergens are from pollens (grass, tree and weeds), pets (cats, dogs, rabbits, horses) and HDM. Specific IgE assays for to all these allergens are available. It is of interest as to how little allergen is needed to evoke a response (the estimates for daily pollen or mite exposure is 5–50 ng/day [48,49] showing the exquisitely sensitive nature of the IgE system. It is easy to advise and promote aeroallergen avoidance, but it is very difficult to achieve or maintain. Physical barriers such as wraparound sunglasses and peaked caps may stigmatize a child at school. HDM and cat are common ‘perennial’ allergens in the United Kingdom and pharmacological treatment may be needed year-round. Nasally inhaled corticosteroids are very effective in most cases of allergic rhinitis, but do not alter the natural history of the condition. Meta-analysis of subcutaneous injected (SCIT) and sublingual immunotherapy (SLIT) for HDM show they are effective treatments in allergic airway disease [50,51]. There have been intriguing studies (of varying design strength) to suggest that allergen (SCIT) monotherapy may prevent the development of sensitization to other aeroallergens and the onset of asthma [52,53]. This consolidates the view of the nose and the lungs constituting the ‘United Airway’, but the therapeutic role of SLIT in children remains uncertain.
It may surprise allergists that the role of allergens in childhood asthma is still debated. In the 1970s, Sarsfield  showed that 85% of children with asthma bad enough to be seen in an out-patient clinic were sensitized to HDM, and Ponsonby  showed clearly that sensitized children are over-represented in the severe range of the asthma spectrum. A case can be made for a causal association ; however, genetic association studies do not show strong links between asthma and IgE loci. Allergen sensitization is a factor that needs to be considered in any asthmatic child. Allergen avoidance measures are difficult in the domestic environment, and their effectiveness seems doubtful .
Humanized anti-IgE (Omalizumab®, Novartis, Frimley, UK) complexes free IgE, decreasing free (but not total) IgE in serum. Its use in difficult adult asthma cases is established  and its cost-effectiveness appears comparable to that of other biological agents used in other chronic conditions . Its use in children remains under review, but there is creeping use of it in difficult paediatric cases (author's observation).
Adverse reactions to antibiotics may be confused with the underlying infective process, as many viruses cause non-specific rashes. Not all the mechanisms of drug reactions are IgE-related and the in vivo and in vitro tests for drug-specific IgE are generally not reliable. The gold (and probably only) standard for investigation of an IgE-mediated reaction is a drug challenge in the hospital environment, but this can be restricted to where an alternate drug is not available and time is available to make the diagnosis. In urgent, serious cases (complicated cystic fibrosis, for example), where no alternative drug is available, it is possible to use a rush desensitization protocol .
This topic was reviewed recently in a paper from the same series . MacDougall  published a retrospective review of death certification and a prospective review through the British Paediatric Surveillance Unit (to which 93% of the UK paediatricians responded) to report deaths and serious anaphylactic reactions. Over a 10-year period there were eight deaths (four to milk: ages 9, 13, 13 and 15 years; two to peanut: 13 and 15 years; one to egg white: 3 months; and one to mixed foods: 5 years). Severe but non-fatal reactions were seen in 55 children over the 2-year prospective period. In US series peanut and soya are more prominent. Troublesome or poorly controlled asthma appears to be a major adverse risk factor for a fatal outcome in anaphylaxis. Using these figures the estimates are that the risk of death in a food-allergic child is 1 in 800 000. These figures have been criticized as representing only the tip of a much larger iceberg because of the use of an over-restrictive definition of anaphylaxis . Clark and Ewan extrapolated from data in a regional allergy centre in Cambridge and estimated the risk of severe reaction to nuts to be 38- to 62-fold higher . The difficulty is that sudden deaths and asthma deaths may be misreported, and there are at least four asthma deaths/million population/annum in the 5–14-year-old age group. There is a clear need for systematic evaluation of allergy status in all respiratory deaths in children and young adults, as the allergic fatalities identified in other ways must surely be only a subset of such deaths .
Intramuscular injection of the correct dose of adrenaline (epinephrine) is the treatment of choice in infants and children with anaphylaxis . This is available in the United Kingdom as the Epipen® or Anapen® (ALK-Abello, Reading, UK) preloaded devices. The Epipen is a self-firing injection containing adrenaline (0·15 mg for children less than 30 kg, Epipen Junior®; 0·3 mg for those greater than 30 kg, Epipen) and the Anapen (and Anapen Junior®) require an active pressing of a button to fire.
For infants weighing less than 10 kg there is a dilemma as to what injection system to use [65,68]. The correct dose is 0·1 ml per kg body weight of a 1:10 000 dilution of adrenaline. Adrenaline drawn up from an ampoule is a possibility, but requires a calmness that may not be present in an emergency situation . Concern about the unavailability of weight-appropriate auto-injectors for babies is offset partially by two factors: the relative unlikelihood of the need for adrenaline in this age group (frequent dietary accidents and anaphylaxis are both uncommon in this age group) and children's physiological capacity to tolerate adrenaline better than adults. Consensus is developing that it is better to treat a baby with slightly too much intramuscular adrenaline on a dose-for-weight basis than to not give any adrenaline. Intravenous use of an inappropriate dose of adrenaline may be lethal .
Who should receive adrenaline?
This decision can either be based on a legal or a medical viewpoint. The former suggests that if a child shows any evidence of sensitivity to an allergen (i.e. a wheal size greater than 3 mm) then one is failing in the care of that child if provision is not made to cover all eventualities, and the provision of injectable adrenaline is mandatory. This is the usual scenario in American practice. Kemp  has challenged this approach by suggesting it is unnecessary and expensive (estimating that it costs $A20 million dollars for each life saved). The medical criteria are more selective, based on the severity of the reaction (any wheeze, laryngospasm, asthma, cough, hypotension, collapse or loss of consciousness require adrenaline). A medical consensus is emerging regarding prescription practices [71,72].
The primary protection of ‘anaphylactic’ children is identification of reactivity followed by strategies to avoid the relevant allergen. Following this is the use of an anti-histamine and perhaps the provision of injectable adrenaline. An act of omission seems more serious than one of widespread prescribing, when the cost is not borne directly by the patient; but there are many Epipens gathering dust in schools and at home and not being carried or used appropriately. New models of wider availability and a move away from provision on a ‘named patient basis only’ should be considered . If adrenaline is prescribed training is essential, and action plans for schools and homes are also necessary.
Stinging insect anaphylaxis
Systemic allergic reactions in children are rare [74,75]. The cooler and wetter northwest European climate possibly makes bees and wasps more docile than in other areas. It is important to remember that stings are meant to be painful and cause local swelling, so only systemic reactions warrant investigation. In addition, the site of the sting can affect the severity of the reaction (the more central the greater the problem appears to be – being stung on the pharynx by a wasp will usually cause more problems than a sting to the fingertip). ImmunoCAP and skin test reagents are available to ascertain sensitization. Lifestyle advice is important (wear dull-coloured clothes (so not to resemble a flower), long sleeves, trousers and shoes, avoid scents and perfumes, use sports bottles with nozzles rather than opaque cans of drinks) and a rescue medication package. Insect-sting anaphylaxis is one of the few indications for immunotherapy that are well recognized in the United Kingdom.
There are groups of children who give surprisingly similar histories of ‘altered reactivity’ and who are referred to an allergy clinic. There are no reliable or proven objective tests for these conditions , although many are offered commercially. Usually these clinical conditions are distinguished easily, and management comprises monitoring improvement clinically during avoidance and deterioration during re-exposure and ensuring that unsubstantiated dietary and lifestyle restrictions are avoided.
Non-IgE food allergy has been described and is often associated with vomiting, often delayed. There has been much recent interest in cow's milk allergy, gastro-oesophageal reflux and eosinophilic oesophagitis , discussion of which is beyond the scope of this paper.
Another group of children referred are those where diet is said to alter behaviour, from hyperactivity to the worsening of autistic symptoms. For the former there seems a cluster of foods (oranges, chocolates, cola) that are recalled as causing problems; for the latter, the ‘neurotoxic’ effects of gluten feature highly. Parents are often desperate for an external cause for what appear to be ‘intrinsic’ disorders. There is no doubt that some children do experience a ‘sugar rush’, or their behaviour worsens with a junk-food diet; however, easy objective tests are lacking. For an individual child's overall management diet may play a small role, but sometimes a very positive one. McCann's controversial findings of an effect of a combination of additives on behaviour in young children  will probably need replication with single additives to gain medical acceptance, although the public already seem to have accepted it wholesale.
The chronic urticarias (of which more elsewhere in this series) are rarely allergen-driven . The physical urticarias (hot, cold, solar, pressure) are occasionally seen and have a clear explanation and can usually be demonstrated. Hereditary angioedema is well described in children, and management does not differ greatly from that in adults .
Much of what paediatric allergists do is simple and straightforward, providing a diagnostic and treatment service for food allergy, asthma, eczema, rhinitis and drug allergy. The major problem is the sheer volume of referrals from primary care and the paucity of clinics at both secondary and tertiary levels. The role of tolerance is intriguing, and how this can be achieved simply and safely using immunotherapy will be a fertile area of research for several years to come.
Sadly, Richard Sporik passed away in March 2008. Richard was a Consultant Paediatrician at Treliske Hospital in Truro. He had a longstanding interest in allergy and developed a top-quality paediatric service for West and Central Cornwall. Richard was intelligent, gentle and modest and was highly thought of by both colleagues and patients. After the diagnosis of his chronic illness he continued to work, one of his last acts being to write this review on paediatric allergy which he almost completed. Richard will be sadly missed by all his colleagues – Edward Kaminski, Derriford Hospital, Plymouth, UK.