Review article: the diagnosis and management of food allergy and food intolerances

Transcription

1 Alimentary Pharmacology and Therapeutics Review article: the diagnosis and management of food allergy and food intolerances J. L. Turnbull*, H. N. Adams & D. A. Gorard *Department of Paediatrics, John Radcliffe Hospital, Headington, Oxford, UK. Department of Gastroenterology, Wycombe Hospital, High Wycombe, Bucks, UK. Correspondence to: Dr D. A. Gorard, Wycombe Hospital, Queen Alexandra Road, High Wycombe, Bucks HP11 2TT, UK. nhs.uk Publication data Submitted 11 May 2014 First decision 23 May 2014 Resubmitted 7 September 2014 Resubmitted 14 September 2014 Accepted 16 September 2014 EV Pub Online 14 October 2014 This commissioned review article was subject to full peer-review and the authors received an honorarium from Wiley, on behalf of AP&T. SUMMARY Background Adverse reactions to food include immune mediated food allergies and nonimmune mediated food intolerances. Food allergies and intolerances are often confused by health professionals, patients and the public. Aim To critically review the data relating to diagnosis and management of food allergy and food intolerance in adults and children. Methods MEDLINE, EMBASE and the Cochrane Database were searched up until May 2014, using search terms related to food allergy and intolerance. Results An estimated one-fifth of the population believe that they have adverse reactions to food. Estimates of true IgE-mediated food allergy vary, but in some countries it may be as prevalent as 4 7% of preschool children. The most common food allergens are cow s milk, egg, peanut, tree nuts, soy, shellfish and finned fish. Reactions vary from urticaria to anaphylaxis and death. Tolerance for many foods including milk and egg develops with age, but is far less likely with peanut allergy. Estimates of IgE-mediated food allergy in adults are closer to 1 2%. Non-IgE-mediated food allergies such as Food Protein-Induced Enterocolitis Syndrome are rarer and predominantly recognised in childhood. Eosinophilic gastrointestinal disorders including eosinophilic oesophagitis are mixed IgE- and non-ige-mediated food allergic conditions, and are improved by dietary exclusions. By contrast food intolerances are nonspecific, and the resultant symptoms resemble other common medically unexplained complaints, often overlapping with symptoms found in functional disorders such as irritable bowel syndrome. Improved dietary treatments for the irritable bowel syndrome have recently been described. Conclusions Food allergies are more common in children, can be life-threatening and are distinct from food intolerances. Food intolerances may pose little risk but since functional disorders are so prevalent, greater efforts to understand adverse effects of foods in functional disorders are warranted. Aliment Pharmacol Ther 2015; 41: doi: /apt.12984

2 J. L. Turnbull et al. INTRODUCTION Although food allergies are commonly encountered by paediatricians, and although the public and lay press demonstrate a marked interest in food allergies and intolerances, this diagnostic arena is little regarded by most gastroenterologists dealing with adult patients. Adverse reactions to foods vary in clinical presentation, severity and underlying aetiology. Patients, the public, doctors and other health professionals frequently confuse non-allergic food reactions with food allergy. Adverse reactions to foods can be broadly divided in to those with an immune basis food allergies and coeliac disease, or those without an immune basis termed food intolerances (Figure 1). Although coeliac disease is a T-cell mediated (type 4 hypersensitivity) immune response to gluten, it is not usually classified as a food allergy and is not discussed further. Acute toxic reactions to food contaminated by bacteria or by aflatoxins, or to food containing excess histamine such as spoilt fish (scombroid food poisoning), are not reviewed here. Furthermore, adverse reactions to foods arising from metabolic errors (usually autosomal-recessively inherited enzyme deficiencies) and culminating in serious disease such as phenylketonuria, tyrosinaemia, organic acidaemias, homocystinuria, Refsum s disease, galactossaemia are not discussed. Malabsorptive problems such as abetalipoproteinaemia and pancreatic insufficiency in which fats aggravate bowel symptoms are similarly not included. Although dietary intervention can influence Crohn s disease 1 3 and orofacial granulomatosis, 4 the role of food in these inflammatory illnesses is also beyond this article s scope. METHODS A literature search was performed, using OVID MED- LINE, EMBASE and the Cochrane library, up until May Search terms included food allergy food intolerance, IgE-mediated, non-ige mediated, cow s milk protein allergy, Protein-induced enterocolitis syndrome, anaphylaxis, immunotherapy, eosinophilic oesophagitis, eosinophilic gastroenteritis, lactose intolerance, fructose intolerance, gluten sensitivity. The articles returned by the search were selected based on English language and relevance to this review. Important articles identified in the most recent published reviews were also then appraised. What is food allergy? Food allergy is an adverse immune-mediated response, which occurs reproducibly on exposure to a given food Adverse reactions to foods Food allergy Immune mediated Non-immune mediated Toxic reactions Food intolerance Toxins Bacterial toxins Aflatoxins Scombroid poisoning Pathophysiology explained unexplained IgE-mediated Urticaria Angioedema Bronchospasm Rhinitis Laryngospasm Diarrhoea/ vomiting Anaphylaxis Oral Allergy Non-IgEmediated Food protein -induced enterocolitis syndrome Food protein -induced proctocolitis Food protein -induced enteropathies Mixed IgE- and non-ige-mediated Cow s Milk Protein Allergy Eosiniphilic Oesophagitis Eosinophilic Gastroenteritis Coeliac Disease Pharmacological caffeine tyramine Enzyme deficiencies lactose/ fructose malabsorption Non specific gut and non-gut reactions to food includes Irritable Bowel Syndrome & other functional GI disorders: -luminal distension -heightened gastrocolic reflex -altered perception Behavioural factors: -expectation/conditioning Psychological factors Figure 1 Classification of adverse reactions to foods. 4 Aliment Pharmacol Ther 2015; 41: 3-25

3 Review: food allergies and intolerances and is absent during avoidance. A diagnosis of food allergy requires evidence of sensitisation and specific symptoms on exposure to a particular food. The immune response in food allergy can be classified into IgE-mediated, non-ige-mediated or a mixture of both (Figure 1). IgE-mediated food allergy requires food allergen sensitisation (with the development of serum specific IgE antibody to a food allergen), and secondly the development of signs and symptoms on exposure to that food. In non-ige-mediated food allergy, T-cell-mediated processes predominate and there may be histological evidence of an underlying immune process such as eosinophilic inflammation of the gastrointestinal tract. What is food intolerance? Other types of undesirable reactions to food are termed food intolerances. These non-allergic food reactions do not involve the immune system. Some food intolerances involve an organic pathophysiological process, e.g. lactose intolerance occurs as a consequence of deficiency in the enzyme that breaks down lactose. However, some food intolerances cannot be readily explained by currently understood organic processes, e.g. many of the food intolerances reported in irritable bowel syndrome (IBS) patients. Prevalence of food allergy and other adverse reactions to food Perceptions of adverse reactions to food, whether allergy or intolerance are common. In a United Kingdom household survey, 20% of the population reported food intolerances. However when double-blind placebo-controlled food challenges were performed, the prevalence of true reactions to food was less than 2%. 5 In a German study, one-third of respondents to a postal questionnaire reported reactions to food, but subsequent double-blind placebo-controlled food challenges identified a true prevalence of adverse food reactions of 3.6%, of which more than two thirds were IgE-mediated. 6 Women were more likely to have both self-reported symptoms and blind challenge-confirmed adverse food reactions. 7 The prevalence of true immunologically mediated food allergy is difficult to measure. There are large variations in study methodology, from those relying on self-reporting questionnaires that tend to vastly exceed the true prevalence due to their reliance on lay perceptions of allergy, to studies using more rigorous double-blind placebo-controlled food challenges but only including small numbers of patients. What is clear is that food allergy is more common in children than adults, and seems to be increasing in prevalence in many countries The prevalence of food allergy across the whole US population is approximately 2.5 3% when objective measures, such as serological testing and food challenges, are used. 11, 12 While prone to overestimation, population surveys of US children yield self-reported prevalences of food allergy that vary from 4% in the 2007 National Health Interview Survey, 13 to 8% in another population survey of US children. 14 More objective serological data from the 2005 National Health and Nutrition Examination Survey, established that 4.2% of US children aged 1 5 years had positive food-specific serum IgE serology results for peanut, milk, egg white or seafood. The prevalence of elevated specific IgE reduced with age to 3.8% in 6 19 year olds, and 1.3% in those over 60 years. 12 Since relatively few epidemiological studies have used the gold standard of diagnosis the double-blind placebo-controlled food challenge, the true prevalence of food allergy remains elusive, and a US-based metaanalysis from 2010 was forced to give a range from 1 10%. 9, 15 The ongoing European Union-commissioned birth cohort study EuroPrevall will provide important epidemiological data on allergy in Europe. 16 IgE-MEDIATED FOOD ALLERGY The very nature of the gut requires that its large mucosal surface is continually vulnerable to foreign substances, from food proteins to commensal bacteria and pathogens. The mucosal immune system is required to mount protective responses against pathogenic organisms and toxins, whilst not responding excessively to harmless commensal bacteria and food components. In health, the physical properties of the gut, including gastric acid, digestive enzymes, mucosal integrity and mucus secretion, reduce the penetration of ingested pathogens and food proteins. The innate immune system targets molecules common to many pathogens. There are also targeted responses from the adaptive immune system to either tolerate or react to specific proteins, utilising lymphocytes, cytokines, secreted IgA and Gut Associated 17, 18 Lymphoid Tissue. This immune exclusion minimises the amount of potentially allergenic protein taken up from the gastrointestinal tract, but nonetheless more than 100 g of immunologically recognisable dietary protein can still be absorbed by those following North American diets. 19 The interaction of this protein load with the gastrointestinal immune system, particularly in early life, is key in determining the response to individual proteins in future. An interplay of factors determines whether there Aliment Pharmacol Ther 2015; 41:

4 J. L. Turnbull et al. will be a mucosal IgA secretion response, a systemic immune response, or local and systemic tolerance upon re-exposure. The mechanisms leading to food protein tolerance are not fully understood but tolerance involves an active regulatory response mediated by regulator T (T Reg) cells, and clonal deletion which removes T cells with undesirable targets. Antigen-presenting macrophages and dendritic cells may be modulated by commensal bacteria, so they more readily secrete cytokines encouraging T Reg cell amplification and tolerance. Current theories behind tolerance and hypersensitivity are more thoroughly detailed by Brandtzaeg. 18 IgE-mediated food allergy results when an adaptive immune response that is effective in targeting parasites is instead mounted against food proteins. Food and other proteins enter the body by ingestion, inhalation and skin penetration. They are taken up by antigen presenting cells and presented to T Helper cells. The resulting cytokine release determines whether there is a predominant T H 1orT H 2 response. The T H 1 response is the cell-mediated response, effective against intracellular bacteria and protozoa. In contrast the T H 2 response targets parasites such as helminths with stimulation of eosinophils, basophils and mast cells, along with IgE-producing B cells. When food proteins inappropriately trigger a T H 2 predominant response the result is Type 1 hypersensitivity and IgE-mediated food allergy. In IgE-mediated allergy, exposure to a trigger food brings about immediate and consistently reproducible effects on the gut, skin or respiratory system. Any protein can theoretically cause sensitisation, but the vast majority of these hypersensitivity reactions occur to the same most common allergens. The common allergen proteins tend to be water-soluble glycopeptides with greater resistance to proteolytic enzymes, acid and heat. In children, cow s milk, egg, peanut, soy, tree nuts, fish, shellfish and wheat account for 85% of all food allergies (Table 1). However, many of these early onset allergies are associated with the eventual development of tolerance and this leads to a different spectrum of allergies in adults. The most common allergens in adults are peanuts, tree nuts and seafood. Tree nuts include pistachio, pecan, Brazil, cashew, hazel and walnut, and many others. Although some individuals are allergic to just one tree nut type, there is much cross-reactivity so individuals tend to be allergic to multiple varieties of tree nut. Symptoms of food allergy usually start within minutes of exposure to the trigger food and must occur within 2 hours to classify as an IgE-mediated response. Any combination of local oral, dermatological, gastrointestinal Table 1 Prevalence of food allergies in adults and children 30 Food Prevalence (%) Young children Cow s milk 2.5 Egg 1.3 Peanut 0.8 Soy 0.4 Tree nut 0.2 Shellfish 0.1 Adults Shellfish 2 Peanut 0.6 Tree nut 0.5 Fish 0.4 and respiratory symptoms can occur (Table 2). The most severe reactions have systemic effects leading to potentially fatal anaphylaxis. Cutaneous effects are the most common, found in 80% of cases. 20, 21 Respiratory symptoms usually occur only as part of a generalised reaction and isolated respiratory symptoms should prompt a reconsideration of the diagnosis of food allergy. Gastrointestinal symptoms can occur through immediate gastrointestinal hypersensitivity. Oral allergy syndrome (pollen food allergy) Oral allergy syndrome, or pollen food allergy, is a localised IgE-mediated food allergy triggered by certain fruits and vegetables. Symptoms affecting the lips, mouth and throat are usually mild without any associated systemic reaction. The phenomenon results from the similarity between epitopes present in fruit and vegetables and in common pollens such as birch and ragweed. Pollensensitised individuals, often with clear symptoms of seasonal rhinitis, can therefore mount an IgE response to structurally similar fruit and vegetable allergens upon oral contact. Their food allergy symptoms tend to be worse during times of high pollen counts. Associated systemic symptoms can occasionally occur, and anaphylaxis has been reported in around 2%. 22 Oral allergy syndrome affects around 2% of the UK population and 50 90% of those with birch pollen allergy. 23 It is more common in adults than children. Exposure to trigger foods causes an immediate contact-related urticaria, with associated itching and tingling of the lips, tongue and throat. Occasionally there are more severe features such as frank angioedema or blistering. Common triggers are melon, banana, apple, kiwi, tomato and celery. Reactions usually occur only with the raw food 6 Aliment Pharmacol Ther 2015; 41: 3-25

5 Review: food allergies and intolerances Table 2 Clinical features of IgE-mediated food allergy Local oral & orbital Dermatological Gastrointestinal Respiratory Systemic Itching of palate/lips Acute urticaria Nausea Nasal itching Hypotension Swelling of lips/ tongue Flushing Abdominal pain Rhinorrhoea & nasal obstruction Eye itching, redness and Angioedema Vomiting Sneezing watering Periorbital oedema Exacerbation of existing Diarrhoea Laryngospasm eczema Morbiliform rash Dyspnoea, wheeze and not to cooked or processed forms, since heating attenuates the culprit allergens. 20 Particular pollen allergies are associated with a corresponding group of food allergies (Table 3). Examples of identified overlapping antigens include the Bet v 1-cross reactive antigens that have a role in the food allergies commonly associated with birch pollen allergy. Bet v 1 comprises a plant defense protein, but there are homologous proteins found in apple (Mal d 1), cherry (Pru av 1) and celery (Api g 1). 24 In contrast, where oral allergy syndrome occurs in the absence of pollen allergy the culprit allergens are more likely nonspecific lipid transfer proteins. These are often concentrated in the peel of, for example, apple (Mal d 3), cherry (Pru av 3) and orange (Cit s 3). 25, 26 IgE-mediated allergy testing can be performed in oral allergy syndrome, preferably using raw foods in skin prick testing. Managing oral allergy syndrome involves avoidance of raw trigger foods alongside antihistamine treatment of seasonal allergic rhinitis. Since many of the more immunogenic proteins are found within a fruit/vegetable s skin, peeling the offending fruit/vegetable may diminish the reaction. Those with solely mild oral symptoms may choose to continue consuming culprit foods. The minority of individuals who experience systemic reactions should strictly avoid trigger foods and carry adrenaline (epinephrine) autoinjector pens. Cross-reactivity also occurs in latex allergy % of individuals who have allergic reactions on contact with latex have an increased likelihood of being allergic to banana, avocado, chestnut, kiwi, tomato, bell pepper and often to some other fruits too. Hevein is a latex protein with cross-reactive epitopes shared by some fruits. Cross reacting epitopes include those of the Hev b class including Hev b 2 27, 28 (beta-1,3-glucanase). Anaphylaxis Anaphylaxis is a systemic response initiated by binding of antigen to membrane-associated IgE on mast cells and basophils. This causes the release of inflammatory mediators, including histamine, tryptase and chemo-attractants. Injected histamine can reproduce most of the features of anaphylaxis in animals, where it leads to smooth muscle spasm, eosinophil activation and increased vascular permeability. 29 Food anaphylaxis can occur after ingestion and also after exposure via other routes such as skin contact with vomit, or inhalation of minute food particles released during cooking. Food anaphylaxis presents differently to drug- or venom-induced anaphylaxis in that the mechanism for fatal reactions is almost always respiratory arrest, with isolated cardiovascular collapse being rare. Gastrointestinal features are more common in food anaphylaxis and occur in 41% episodes, compared to 3.7% of anaphylaxis episodes induced by drugs/venoms. 29 Biphasic or protracted reactions are more common in food anaphylaxis. 30 Table 3 Food groups cross reacting with pollens in Oral Allergy Syndrome (adapted from ref 22) Pollen Allergen Associated cross-reacting Oral Allergy Syndrome triggers Birch Bet v 1 Apple, peach, plum, cherry, apricot, almond, carrot, celery, parsley, hazelnut (also possible soy and peanut systemic allergy) Ragweed amb a allergen group Melon, cucumber, zucchini, banana, kiwi Mugwort art v 1 Celery, carrot, parsley, peppers, mustard, cauliflower, broccoli, garlic, onion Orchard Grass dac g allergen group Melon, peanut, potato, tomato Timothy Grass Phl p allergen group Swiss chard, orange Aliment Pharmacol Ther 2015; 41:

6 J. L. Turnbull et al. Sampson et al. have drawn up three sets of clinical criteria for the diagnosis of food allergy-induced anaphylaxis (Table 4). 30 An elevated serum tryptase soon after a suspected reaction can identify anaphylaxis. However tryptase levels are less commonly elevated in anaphylaxis when it is caused by food than when triggered by drugs or venoms. This may be because basophils play more of a role than mast cells in food anaphylaxis. Although there are increasing numbers of emergency hospital attendances for food-induced anaphylaxis 31 deaths from food-induced anaphylaxis are uncommon. A meta-analysis concluded that the incidence of death due to food-induced anaphylaxis was 1.8 per million person-years in food-allergic individuals. 32 In these fatalities, peanut is the most common food allergen, and there is invariably a history of asthma. Food dependent exercise-induced anaphylaxis Exercise-induced anaphylaxis is a rare disorder affecting adults and children, in which allergic symptoms occur during or after exercise. Symptoms range from urticaria, angioedema, respiratory and gastrointestinal signs to anaphylactic shock. 33 In one-third of patients there is an association with certain food allergens, 34 and the condition is then termed food dependent exercise-induced anaphylaxis (FDEIA). In FDEIA, allergic symptoms typically occur two hours after ingestion of allergen food and with subsequent exercise. 35 FDEIA may also rarely occur if the food is ingested soon after the completion of exercise. 36 Shellfish and wheat are the most common causative foods in both children and adults. Alcohol, tomatoes, cheese, peanuts and celery are also often implicated. Food processing and the presence of 33, 34, 37, 38 undigested protein allergens in the digestive tract at the time of exercise may be a prerequisite for the development of FDEIA. This has been suggested after studying a patient with FDEIA who could safely exercise after soy milk, but not after tofu a coagulated form of soy milk that is more slowly digested. 39 Other factors linked to FDEIA include nonsteroidal anti-inflammatory drugs, 40, 41 a history of atopy and consumption of alcohol with the food allergen. 41 The pathophysiological mechanisms leading to a temporary loss of immune tolerance in FDEIA have not been established. The Transient Receptor Potential channel, TRPV1, functions as an ion channel in neurones. TRPV1 regulates the cellular influx of calcium, and can be activated by heat, acidosis and various chemical stimuli. Mast cells are closely anatomically related to neurones bearing TRPV1 receptors, and anaphylaxis is associated with TRPV1 activation. Exercise increases temperature and metabolic acidosis which lower the threshold for TRPV1 activation, and may provide an explanation for exercise-induced anaphylaxis. 42 Furthermore exercise increases gut permeability, 43 and food allergens may more easily permeate the gut wall during exercise gaining greater access to the gut-associated immune system. 44 NSAIDs and alcohol also increase gut permeability, explaining why they can exacerbate 41, 45 FDEIA. Additionally, exercise diverts blood from inactive tissue to active tissue. Food-sensitised, gut-associated immune cells may redistribute to the skin and skeletal muscles during exercise. The increased amount of food allergen presented to the mast cells and basophils precipitates histamine release and the allergic reaction of FDEIA. 46 The diagnosis of FDEIA is extremely challenging since prior food allergy may not be suspected and food challenge in the absence of exercise is usually negative. Once a responsible food allergen is identified, it should be Table 4 Diagnostic criteria for anaphylaxis (adapted from ref 30) Anaphylaxis is highly likely when any 1 of the following three criteria sets is fulfilled: 1. Onset of illness in minutes to several hours with involvement of the skin or mucosa (urticaria, itching or flushing, swollen lips-tongue-uvula) and one of: a Respiratory compromise (dyspnoea, wheeze, stridor, reduced peak expiratory flow, hypoxia) b Reduced blood pressure or associated symptoms of end organ dysfunction (eg syncope, incontinence) 2. Known allergy and likely exposure to allergen for that patient and with rapid symptoms including two of: a Skin/mucosa involvement (urticarial, itching/flushing, swollen lips-tongue-uvula) b Respiratory compromise (dyspnoea, wheeze, stridor, hypoxia) c Reduced blood pressure or associated symptoms (collapse, syncope, incontinence) d Persistent gastrointestinal symptoms (cramping, abdominal pain, vomiting) 3. Reduced blood pressure after exposure to known allergen for the patient: Adult: systolic pressure <90 mmhg or >30% lower than patient s baseline Child: refer to age-specific centile charts 8 Aliment Pharmacol Ther 2015; 41: 3-25

7 Review: food allergies and intolerances avoided for at least 4 6 h before exercise and also for a 33, 36 period after exercise. If the food allergen is not identifiable then avoiding all foods, nonsteroidal anti-inflammatory drugs, aspirin and alcoholic beverages for 4 6 h before and a period after exercise is advisable. Prophylactic agents such as cetirizine combined with montelukast, 47 and mast cell degranulation inhibitors such as sodium cromoglycate 48 and ketotifen 49 may provide some protection. Exercise should be terminated immediately if allergic symptoms develop. Diagnosis of IgE-mediated food allergy In confirming a diagnosis of a food allergy, the history is key, with tests playing a supportive role. Performing allergy tests and interpreting their results in the context of the history, requires specialist knowledge. Furthermore in vivo allergy tests can be hazardous. Thus in most cases, a diagnosis of food allergy is best made by a specialist clinic. However, a recent UK guideline is recommending primary care based diagnosis for cow s milk allergy in children. 21 A detailed history is essential, but is not sufficient alone to diagnose food allergy. 9 The goal of the history is to identify a likely cause for symptoms, decide whether any allergy is likely to be IgE-mediated and then guide appropriate testing of this allergen and its likely cross-reactive foods. It is important to be focussed with testing, as most tests in allergy are sensitive but not specific. Therefore indiscriminately testing for large batteries of allergens will yield many false positive results. It is important to consider reactions to cooked and uncooked forms of food. 9, 21 The history should also address any personal and family history of asthma, eczema or allergic rhinitis, and family history of food allergies. There is a strong link between atopic eczema and food allergy. The development of atopic eczema before 6 months of age, and severe eczema within the first year of life are associated with the development of egg, milk and peanut allergies. 50 Finally the response to any dietary restrictions and medications should be explored. Testing for IgE-mediated food allergy. Approved tests of value in diagnosing IgE-mediated food allergy are the skin prick test, and measuring food-specific IgE antibody levels. The double-blind placebo-controlled food challenge remains the gold standard in food allergy testing, but is not frequently performed due to its inherent risks, inconvenience and cost. Alternative tests, some of which are readily available to the public, are specifically not recommended. These include Vega (electrodermal) testing, hair analysis, applied kinesiology, serum-specific IgG4, lymphocyte stimulation, facial thermography, gastric juice analysis, endoscopic allergen provocation, cytotoxicity assays and the Mediator Release Test â. In addition atopy patch tests are not helpful in the diagnosis of IgE-mediated food allergy, 9, 21 but may have a role in testing for T cell mediated immune responses. The main difficulty with both specific IgE (sige) testing and skin prick tests is that they cannot distinguish between an individual who is merely sensitised to the allergen (has detectable circulating sige) and one who is clinically allergic (has mast cell-bound IgE leading to immediate mediator release). This is why it is essential to target allergy testing based on leads in the patient s history, and it is ideal to complement the tests with a definitive food challenge. 51 Skin prick tests are usually first line, but sige tests are preferable where there is a significant anaphylaxis risk, in pregnancy, in severe skin disease and dermatographism and in patients unable to stop taking medications such as b-blockers and antihistamines. Food allergen specific serum IgE. The measurement of sige antibodies in serum was originally done by radioallergosorbent test (RAST â ) assay. However, more accurate methods including fluorescence enzyme-labelled tests are available for measuring sige. Thus, other assays such as ImmunoCap â, Immunlite â and HYTEC-288 â systems are now in use. 51 The assays involve a surface-fixed allergen which is incubated with the patient s serum. Any sige antibody to the antigen will bind to the fixed allergen and remain after washing. Bound IgE is then identified by the binding of labelled anti-ige. The methods in common use are all reliable, but require different reference ranges for clinical decision points. 51 The level of sige correlates with the likelihood of a clinically significant reaction to that food trigger, but does not indicate its likely severity. Sampson et al. originally drew up values of sige to the most common allergens that had a 95% positive predictive value in the diagnosis of food allergy, based on a study of more than 100 children where tests were correlated with the gold standard food challenge (Table 5). 52 SIgEs are not helpful if the history is not supportive of IgE-mediated allergy as there is a reasonably high chance of a false positive result. Conversely, sige testing cannot rule out allergy where the history is suggestive since its sensitivity is low and between 10% and 25% of significant reactions, including anaphylaxis, may be missed. Individuals with a strong history for IgE-mediated allergy and negative sige testing should undergo an oral food challenge. 53 Aliment Pharmacol Ther 2015; 41:

8 J. L. Turnbull et al. Table 5 Predictive value of food allergen-specific IgE levels (from ref 52) Allergen sige (ku/l) Positive predictive value (%) Egg 7 98 Milk Peanut Fish Tree nuts Soybean Wheat Egg <2 years 2 95 Milk <2 years 5 95 How to use sige tests has recently been elaborated by Steifel and Roberts. 51 They describe a clinical tool that uses history and sige levels to categorise patients into has allergy, probably allergy, possible allergy and no allergy. Patients with a high likelihood of allergy based on history and sige can be diagnosed in the absence of a formal food challenge, but food challenges may be indicated in those with possible allergy (Figure 2). As well as having a role in allergy diagnosis, sige tests can also be useful in assessing whether tolerance has developed in a child previously allergic to egg or milk. With egg, a 50% reduction in sige level has been associated with a 50% chance of tolerating egg at challenge. 54 Advantages of sige tests over skin prick testing include their freedom from any risk of anaphylaxis and their accessibility to nonspecialists. The development of more sige tests against specific protein components is a promising advance. For example in milk allergy, sige to casein rather than to whole milk extract is more specific, with a clinical decision point of 6.6 ku/l being 100% specific. 55 Hong et al. have described the use of component-resolved diagnostics to see whether testing for specific peanut allergens is more predictive of clinical allergy rather than mere sensitisation, and whether this can predict those at risk of anaphylactic reactions. 56 They found that children with clinical allergy had significantly higher IgE reactivity to the peanut allergens Arah h 1, 2 and 3 than asymptomatic but sensitised children. SIgE to Ara h2 was the most discriminatory and a cut-off point of 0.65 ng/ml gave 99.1% sensitivity, 98.6% specificity and a 1.2% misclassification rate in predicting peanut allergy. 56 This compares favourably with using IgE to whole peanut extract (P = 0.008). However, the tests did not identify those at risk for anaphylaxis. Skin prick tests. Skin prick tests also detect the presence of circulating IgE to specific antigens, but are in vivo tests and can provide near-instant results. Food extracts, and controls, are applied to the skin before it is pricked by a lancet. After 20 min the diameter of any induration is measured and compared with tables of clinically significant wheal size for corresponding allergens. Skin prick tests can also be performed for suspected allergies where no commercial tests or extracts are available. In fact for many foods, particularly fruit and vegetables and milk, it is preferable to use real food brought to clinic by the patient, as commercial extracts degrade and lack sensitivity. 57 As with sige, the test cannot distinguish between sensitisation and true allergy, but there is a correlation between the size of wheal and the likelihood of true allergy. The size of the wheal generated varies both with the candidate allergen and the population under study, but there have been attempts at standardising significant decision points for each allergen. 58 Verstege et al. found that the sensitivity and specificity of skin prick tests were relatively poor if using a simple Likelihood of allergy from specific IgE (ku/l) LOW INTERMEDIATE (eg. nut<0.35) (eg. nut 0.35 to <15) HIGH (eg. nut>15) Likelihood of allergy from history HIGH eg urticaria & wheeze on 2 exposures INTERMEDIATE eg urticaria on single exposure LOW eg non-ige symptoms Possible allergy Possible allergy No allergy Probable allergy Possible allergy Possible allergy Allergy Probable allergy Possible allergy Figure 2 Schema for considering diagnosis of food allergy in children and teenagers using specific IgE testing (Steifel & Roberts ) 10 Aliment Pharmacol Ther 2015; 41: 3-25

9 Review: food allergies and intolerances positive or negative result, but by correlating the size of wheal with the results of a double-blind placebo-controlled food challenge they were able to identify clinical decision points with 95% specificity for hen s egg and cow s milk. 58 Atopy patch testing. Atopy patch tests detect delayed hypersensitivity reactions and can also show immediate urticarial reactions. Patch testing involves applying food extract directly to skin on a patient s back and then assessing for erythema, infiltration and papules after h. There is currently no official role for the atopy patch test in IgE-mediated food allergy, but there is some support that it may reflect late phase clinical reactions. It may yet prove to be useful in combination with sige and skin prick tests in avoiding the need for food challenge, but more work is required to standardise such 9, 59 tests. Elimination diets. Although vital in managing allergy, there is no official role for elimination diets in the diagnosis of IgE-mediated food allergy. However, in practice the response to an elimination diet gives useful information in cases where a food challenge is not being performed. Elimination diets do have a diagnostic role in non-ige mediated allergy where there are no specific tests. In these cases, a dietitian should provide advice on a strict elimination diet and there should be a rechallenge after a number of weeks. Double-blind placebo-controlled food challenge. The gold standard test for food allergy is the double-blind placebo-controlled food challenge. This involves a protocol of ingesting increasing amounts of solution containing either the suspected allergen or a placebo over the course of a day whilst monitoring for symptoms. The test is then repeated after a few days using whichever solution was not used first time. Both patient and clinician is blinded to which is a placebo. Food challenges carry a significant risk of anaphylaxis and should therefore only be performed where there are facilities to treat anaphylaxis. In a study of 225 children undergoing peanut allergy food challenges, 48% had a positive reaction and 11% had a severe reaction. Positive reactions were more likely in males with positive sige to peanut and the presence of another food allergy. 60 SIgEs that are able to identify those who can avoid food challenges are therefore a current area of active research. For practical reasons of cost, time-efficiency and difficulties obtaining suitable placebos, open food challenges are often performed instead of blinded challenges. This may be adequate for obvious immediate reactions, but delayed effect reactions should be followed by a full double-blind placebo-controlled food challenge. 61 Food challenges are also used in testing for the development of tolerance to milk and egg with age. Management of IgE-mediated food allergy Managing IgE-mediated food allergy is a two-pronged approach of avoidance and preparation for accidental exposure. 62 There is a vast amount of information to impart to the patient and/or their family, usually beyond the scope of the typical outpatient appointment. The contribution of the multidisciplinary team as well as written information and emergency action plans is invaluable. Avoidance. A list of primary foods to be avoided should have arisen from history taking and appropriate testing for candidate allergens. Likely cross-sensitisation needs to be considered. There are known allergic associations, thought likely due to common epitopes, for instance between peanuts, tree nuts and sesame, and between egg and tree nuts (Table 6). 63 Whether or not to advise avoiding potential cross-sensitising foods must be tailored to the patient, taking into account their level of risk, presence of asthma, and the nutritional implications of extra restriction. The final list of foods to eliminate needs to be expanded to include other foods which may contain these primary ingredients. The allergic patient and family will need advice on how to check food labels, a list of alternative terms used for their allergen (such as casein for milk), and advice on whether the child may be able to tolerate baked forms or processed forms of the food. Table 6 Food allergen cross-reactivities (from ref 63) Primary allergen Cross-reactivities Cow s milk Goat s milk 90%, beef 10% Hen s egg Turkey, duck, goose eggs Soy Cross-reactivities with other legumes uncommon Peanut Other legumes generally well-tolerated Fish Significant cross-reactivity between species Tree nut High cross-reactivity with other tree nuts Crustacean shellfish Significant cross-reactivity between crustaceans, cross-reactivity with molluscs less well defined Wheat Uncommon Aliment Pharmacol Ther 2015; 41:

10 J. L. Turnbull et al. The role of dietitians in this is key. Since elimination diets may be restrictive in one or more nutrients, dieticians can suggest replacement foods or supplements accordingly. Patients often benefit from registering with an allergy organisation, such as anaphylaxis.org, which can be a useful source for reference. Some highly sensitive individuals readily react to contamination, either orally or mucosally. Allergens such as peanut persist in the environment. Peanut antigen was found to be still detectable 110 days after peanut butter was smeared onto a table. 64 Reassuringly, it became undetectable after cleaning with a disinfectant wipe, and this endorses advice to wipe eating surfaces before meals. Adherence to elimination diets is necessarily strict and places significant stresses on patients and their families. Vigilance and anxiety about accidental exposure often lead to restriction of activities and reduced quality of life. 65 One study found that children with peanut allergy were restricted from attending school (10%) or school trips (59%), avoided parties (68%) and restaurants (11%), and many were prevented from playing at friends houses (14%) and attending sleepovers (26%). 66 Travel abroad is particularly challenging. Translated information cards on both dietary requirements and emergency treatment are available and advisable. 67 Allergy education reduces harm but also reduces quality of life, perhaps as patients are not afraid of what they don t know. 68 Knowledge is better in families where a child has suffered anaphylaxis and in those who have joined an allergy organisation. 69 Allergy education is more effective in a multi-disciplinary allergy clinic with a paediatric allergist, dietitian and nurse specialist. 70 Preparation for exposure. Patients and their families should be aware that despite their careful measures they are at risk of an accidental exposure. Very small amounts of the food may be sufficient to cause a severe reaction. Significant reactions during food challenges occur with as little as 1 mg peanut, 1 mg egg, 0.02 ml milk, 5 mg fish and 1 mg mustard. 71 Accidental ingestions of culprit allergens are common. This may be due to incorrect assumptions about contents, misreading of labels and being given food by another adult. One study following peanut allergic children found that by 3 years 50% had been accidentally exposed, and by 5 years 75% had been exposed. 72 Accidental exposures were more likely to occur with foods prepared outside the home, in restaurants and with desserts. Younger children were particularly at risk whilst at children s parties and in the care of extended family or friends. Another retrospective study in peanut allergic children found inadvertent peanut ingestion rates of 14% per year. 73 A review of fatal cases found that peanuts were responsible for 62% of food anaphylaxis deaths, with tree nuts causing 30% of deaths and the rest due to shellfish, fish, milk, eggs and fruit. Restaurants and educational settings were the most common locations of fatal anaphylaxis; invariably the allergic victim was not aware that their allergen was an ingredient in the food they ate. Risk factors for fatal anaphylaxis are known food allergy, asthma, delay or failure in using adrenaline (epinephrine), teenager/young adult, history of a previous severe reaction and failure to recognise symptoms as anaphylaxis. 30 Asthma is common to almost all fatal cases of anaphylaxis. 74 New cases of anaphylaxis should be referred for allergy testing, as the trigger may not be what was initially suspected. In preparation for a likely accidental exposure, patients should be provided with an alert bracelet, an adrenaline autoinjector and an exposure action plan describing features of both mild and worsening reactions, and how to inject adrenaline. 9, 74 In the case of children, all regular care givers should be trained and confident in the use of the autoinjector. The autoinjector should always accompany the child and be regularly checked for expiry. Delay and failure to inject adrenaline is a real concern. A survey of Scottish teenagers found that in 18 reactions requiring adrenaline, it was given in only 11 cases. 75 Skill and confidence in using autoinjectors deteriorates over time and there may be a role for more user-friendly devices with audioinstructions. Antihistamines have a supportive role in treating anaphylaxis. They reduce itching and urticaria but are not lifesaving and should not delay adrenaline injection. Intravenous injection of antihistamines can cause hypotension. Glucocorticoids are similarly not lifesaving but their delayed action may reduce biphasic and prolonged reactions. 74 Full emergency management of acute anaphylaxis is discussed elsewhere. 74 New treatments for IgE-mediated food allergy. Although current food allergy management relies on adherence to a strict elimination diet and prompt response to any accidental exposure, there are some promising leads towards new therapies. One approach aims to induce tolerance to specific allergens using immunotherapy, and another aims to tackle allergies more broadly by general suppression of IgE responses. Immunotherapy uses graded exposure to an allergen to promote tolerance to it. Most studies have assessed 12 Aliment Pharmacol Ther 2015; 41: 3-25

11 Review: food allergies and intolerances oral immunotherapy, but the use of the sublingual route is also under investigation. Most protocols have three phases. The initial escalation occurs in a clinical setting, and is followed by a slower, typically biweekly, escalation of doses in the community. This is followed by consumption of a maintenance dose for a number of weeks, or up to a year in some studies. After the maintenance phase an oral food challenge is performed. Key studies and features of oral immunotherapy are detailed in a review by Nadeau. 76 Although oral immunotherapy can reduce the severity of reactions to accidental exposures in most patients and may provide a potential cure from food allergy in some, there are concerns about safety due to high numbers of reactions during the escalation phase. Immunotherapy in peanut allergy is outlined since peanut is the main allergen in food anaphylaxis deaths. Varshney et al randomised peanut allergic patients to oral immunotherapy or placebo. 77 The patients assigned to oral immunotherapy who completed the protocol all passed a blinded food challenge at 12 months. Alongside this they displayed a decrease in positive skin prick tests, a transient rise and then fall in peanut-sige. In a recent and much larger trial, 62% of peanut allergic children were successfully desensitised using oral immunotherapy. 78 However, in other studies up to 20% of patients cannot reach maintenance doses due to severity of gastrointestinal symptoms. 76 Therefore, despite the promising data in the recent peanut allergy desensitisation trial, 78 the current consensus is that oral immunotherapy has not yet been shown to be an effective safe treatment Immunotherapy can only tackle single allergies at a time. As multiple food allergies are relatively common, a broader suppression of reactions has been attempted using recombinant monoclonal antibodies against IgE. Omalizumab, a recombinant antibody that binds to the C epsilon 3 domain of IgE, is currently licenced for use in severe steroid-dependent asthma with aeroallergen sensitivity. Its potential role in treating food allergy is currently under investigation. It may also have a role as an adjunctive treatment in the escalation phase of immunotherapy. 82 Re-introducing foods into the diet. Most children with allergy to egg and milk will eventually develop tolerance. This is also likely with soy, wheat, and many fruits, vegetables and seeds, though these are less well studied. 83 Tolerance often develops step-wise, with cooked forms tolerated before raw equivalents. It has been found that up to 70% of children with severe reactions to na ıve egg or milk protein can tolerate normal serving sizes of the cooked protein. There is an association between this tolerance of baked products and the presence of antibodies against conformational, and not linear, epitopes. This presence of antibodies against conformational epitopes is also associated with higher rates of eventual tolerance There are benefits of allowing cooked products into the diet as soon as tolerated. These include improved quality of life, easier attainment of adequate nutrition and the potential to accelerate the development of tolerance. For example, regular ingestion of baked milk increases the likelihood of becoming tolerant to raw milk and this corresponds to changes in casein-sige and IgG4. 86 However, there are some concerns from some studies that such exposures may increase the risk of eosinophilic enteropathies. In deciding when to recommence a child on cooked products use of particular siges can be helpful. Thus high sige levels or positive skin prick test to ovomucoid can identify children at greater risk of reaction to cooked egg 87 NON-IgE-MEDIATED FOOD ALLERGY There are three main clinical entities of non-ige mediated allergy. These are Food Protein-Induced Enterocolitis Syndrome (FPIES), Food Protein-Induced Proctocolitis (FPIP) and the Food Protein-Induced Enteropathies. The immune mechanisms underlying these conditions are not well understood, though TNFa has been heavily implicated. A recent study has also found a possible link with the T H 2 response found in IgE-mediated allergy, in that whilst there is no production of IgE, T cells are skewed towards a T H 2 response. High levels of IL-13 and TNFa may be key drivers of intestinal epithelial cell damage and eosinophil infiltration, working through activation of the tumour necrosis factor-like weak inducer of apoptosis fibroblast growth factor-inducible molecule 14 (TEAK-Fn14) axis. 88 Non-IgE-mediated food allergy is almost always confined to childhood, and is less well recognised in adults. 89 Food protein-induced enterocolitis syndrome Food protein-induced enterocolitis syndrome is an uncommon food allergy that causes a distinctive and often dramatic reaction to food. It causes solely gastrointestinal symptoms - vomiting with or without diarrhoea. However these symptoms can suddenly worsen, mimicking acute illness such as sepsis or acute abdominal problems. 90 Consequently it is under recognised. The underlying pathophysiology is not well defined, but there Aliment Pharmacol Ther 2015; 41:

12 J. L. Turnbull et al. is thought to be a key role for stimulation of mucosal T-cells leading to a delayed cell-mediated immune response. TNF-a and a relative lack of expression of TNF-b have been implicated. 91 The resulting inflammation is thought to increase mucosal permeability and lead to rapid fluid shifts that explain the clinical effects. Food protein-induced enterocolitis syndrome is a disease of babies and young children, as most children become tolerant by 3 years of age. An Israeli birth cohort study found an estimated prevalence of 0.34%. 92 Symptom onset could be as early as a few weeks of age, but may be delayed to around 5 months of age in breast-fed babies. 92 The three most common foods causing FPIES are cow s milk, soy and rice. Rarer triggers include oat and other cereals, a range of orange vegetables (sweet potato, squash, carrot), egg white, legumes (peanut, green pea, string bean), chicken, turkey, fish and banana. 93 The onset of FPIES symptoms is typically 2 h after exposure, but can range from 0.4 to 4 h. 91 The presentation is with repetitive forceful vomiting, with or without diarrhoea. 91, 94 Only the gastrointestinal tract is involved, though fluid depletion can give rise to features of shock with lethargy, pallor and cyanosis. 90 Although symptoms resolve within 6 12 h, 91 these children can present acutely unwell. These patients are frequently treated for sepsis, pyloric stenosis or inherited metabolic disease, due both to their severe compromise and the absence of features that usually evoke consideration of an allergic cause such as very recent ingestion and cutaneous or respiratory phenomena. Furthermore test results in FPIES may show metabolic acidosis, neutrophilia and thrombocytosis. 93 The diagnosis is a clinical one, though the presence of blood, eosinophils, lymphocytes or increased reducing substances in the stools is supportive. 94 A study of 66 patients found that it takes up to five FPIES episodes to establish a diagnosis, equating to an 8 month delay in diagnosis. 91 Clues to the possibility of FPIES include a more rapid recovery than would be expected in sepsis, infective gastroenteritis or surgical conditions, and being at a stage of dietary transition. The criteria for diagnosing FPIES (Table 7) are (i) age under 9 months old at diagnosis; (ii) repeated exposure to the incriminated food elicits repetitive vomiting and/ or diarrhoea within 24 h; (iii) solely gastrointestinal symptoms; (iv) dietary elimination of the offending food protein resolves symptoms within 24 h. Infants with FPIES are at increased risk of a subsequent IgE-mediated food allergy. 94 Management of FPIES is allergen avoidance, with associated education and dietetic support. In breast-fed infants, the mother should avoid the offending food only if there has been a documented reaction after maternal consumption. 95 Formula-fed infants will require extensively hydrolysed formulas, with amino acid formulas used only if reactions persist. Accidental allergen exposure will require intravenous or oral rehydration, and in some cases, inotropic support. Steroids may be considered in moderate to severe cases, though there is no evidence that they hasten recovery. 93 There is wide variation in reported rates of resolution of FPIES. In a study of 44 infants with FPIES to cow s milk, 50% developed tolerance by 1 year and 90% developed tolerance by 3 years. 92 Solid food FPIES appears to persist for longer. 96 After months a supervised food challenge should be performed, provided prompt treatment for shock and anaphylaxis is available and intravenous access is already established in those with a 91, 93 history of hypotension. Food protein-induced proctocolitis Food protein-induced proctocolitis is a benign transient condition of otherwise well neonates, who pass blood and mucus in the stool. Cow s milk protein is the most common food culprit. Since it is present in breast milk, Table 7 Diagnostic features of non-ige mediated food allergic conditions Food protein-induced enterocolitis syndrome Food protein-induced proctocolitis Food protein-induced enteropathies Age <9 months at diagnosis Predictable repetitive vomiting and/ or diarrhoea within 24 h of exposure to trigger food Only gastrointestinal symptoms Resolution of symptoms within 24 h once exposure avoided Small and fresh rectal bleeding with mucus in a relatively healthy neonate Absence of systemic symptoms and no other cause found Disappearance of bleeding after diet change, and reappearance after re-introduction Chronic diarrhoea/steatorrhoea Faltering growth 14 Aliment Pharmacol Ther 2015; 41: 3-25