SEARCH

SEARCH BY CITATION

Keywords:

  • food hypersensitivity;
  • immunological tolerance;
  • sensitization

Abstract

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

It is an immunological paradigm that avoidance of food allergen may reduce the risk or prevent immunological reactions and conversely that a greater exposure increases the magnitude of the immune response. Consequently, food allergen avoidance has been recommended to reduce the risk of sensitization in infants and to prevent clinical reactions in children with positive skin prick tests (SPT). In the latter setting, it is hoped that avoidance may either promote or at least not retard the development of tolerance. Animal studies, however, have demonstrated that tolerance to food allergens may be induced by either large (high zone tolerance) or small (low zone tolerance) doses, whereas doses in between may actually stimulate immune responses. In this review, we discuss whether strict allergen avoidance is always the most appropriate strategy for preventing or managing IgE-mediated food allergy.

Foods contain multiple antigens, which stimulate the immune system. In a majority of individuals, immunological reactions are controlled and fail to provoke symptoms despite ongoing antigen exposure, frequently in large (gram) quantities. All individuals mount immune responses to ubiquitous food antigens (reflected in measurable IgG or T cell responses in vitro). The factors leading towards diversion of this normal response to a potentially pathological IgE response in the atopic individual are not well understood. It has been widely held that strict avoidance of food allergens may reduce the risk or prevent the IgE-mediated sensitization and immunological reactions. In this review, we examine whether this is necessarily the most appropriate strategy or not.

Historical recommendations

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

There is evidence for a rising prevalence of food allergy in the latter part of the twentieth century. From the 1980s, it has been considered that avoidance of food allergens in the early life may reduce the risk of food allergy and there have been many recommendations aimed at reducing food allergen exposure by using hypoallergenic formula or delayed introduction of allergenic foods. (Table 1) Such recommendations have not been associated with a fall in food allergy, but rather the last two decades have seen significant increases, in western countries, in the incidence of childhood food sensitization to allergens such as peanut (1), specialist consultations for food allergy (2) and hospitalization for food-related anaphylaxis (3). It has also been suggested that once food allergen sensitization occurs, strict avoidance is necessary to reduce further sensitization and to facilitate the development of tolerance. This suggestion arose from the observation, that food allergic children who had inadvertent exposure giving rise to clinical reactions were less likely to outgrow their food allergy. This conclusion is confounded by the possibility that such children were more sensitive initially and thus subsequently reacted to small amounts of food, which may not have resulted in clinical reactions in less sensitive subjects (4).

Table 1.   Recommendations for avoidance or delayed introduction of allergenic foods
YearPublication typeTitleKey recommendations
1982Review (5)Prediction and prevention of atopic allergyExclusive breast feeding (BF) for 6 months. Casein hydrolysate if unable to BF. Introduction of solids after 4–6 months.
1984Review (6)Prevention of food allergiesBF. Avoidance of allergenic foods in 3rd trimester, lactation and a restricted infant diet.
1988Review (7) Allergy to cow’s milk in the first year of life and its preventionDecreased milk intake during pregnancy. Maternal avoidance of Cows milk (CM) whilst BF. Soy or hydrolysed formulas. Avoid all CM products in infant’s diet in first year.
1990Review (8) Prevention of childhood allergy by dietary manipulationExclusive BF for 6 months. Total avoidance of all CM products and eggs during lactation. Soy milk if not BF. Avoid solids until >6 months, CM products >6 months. Further postpone eggs, fish and peanuts.
1999Joint position statement ESPACI-ESPGHAN (9)Dietary products used in infants for treatment and prevention of food allergy.Exclusive BF for 4–6 month. Inconclusive evidence for maternal dietary avoidance during lactation. Exclusive feeding with hypoallergenic formula if not BF. Avoid solid foods until >5 months.
2000Position statement AAP (10)Hypoallergenic FormulasBF for 12 months or longer. Eliminate peanuts, tree nuts and consider elimination of eggs, fish, CM and perhaps other foods during lactation. Hypoallergenic formula (extensively hydrolysed or amino acid) if not BF. Avoid all solid foods until >6 months, CM products >12 months, eggs >24 months, peanuts, nuts and fish > 36 months.
2004Literature review European AACI (11) Dietary prevention of allergic diseases in infants and small children.Exclusive BF for 4–6 months. No evidence for exclusion diet in lactation. Extensively hydrolysed formula and avoidance of solid foods for 4–6 months. No evidence for the preventive effect of a diet after age 4–6 months (although needs further investigation)
2006Cochrane systematic review (12)Maternal dietary antigen avoidance during pregnancy and/or lactation, or both, for preventing or treating atopic disease in the child.The evidence is insufficient to infer that antigen avoidance during lactation is beneficial or harmful.
2006Cochrane systematic review (13)Formulas containing hydrolysed protein for prevention of allergy and food protein intolerance in infants.Limited evidence that hydrolysed formula is better than cow’s milk formula.
2007Position statement ASCIA (14)Primary allergy prevention in childrenFood avoidance in pregnancy or lactation is not recommended. Delay solids until 4 to 6 months. Avoidance of peanuts, tree nuts and shellfish for 2–4 yr may be recommended in high risk children, (but there is no evidence to support this).

Epidemiology

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

Examination of the epidemiology of food allergy in different countries indicates that; (i) there is a wide geographical variation in prevalence that is not necessarily explained by differing allergen exposure, (ii) an increased exposure in early life is not necessarily related to increased sensitization and (iii) equivalent sensitization in early life may result in a differing prevalence of food allergy in later life.

The prevalence of allergy to frequently ingested common food allergens appears to vary widely even in Europe (15) and is not clearly related to levels of consumption. The prevalence of clinical allergic reactions to peanuts in the UK (1.5%) (1) and North America (1.5%) (16) is much higher than in Israel (0.04%) and whereas per capita peanut consumption is greater in the USA (17.13 gm/capita/day) and it is similar in the UK(8.39 gm/capita/day) and Israel (7.3 gm/capita/day) (17). Despite relatively high levels of peanut consumption in Japan (4.4 gm/capita/day) and Korea (4.15 gm/kg/day), peanut allergy is a relatively uncommon cause of food allergy when compared with other foods such as egg and milk (18, 19). Fig. 1 shows the relative proportions of egg, peanut and milk allergy in community-based populations in different countries. Although the method of collection of information differs (questionnaires, skin prick testing or challenge proven), the figure reveals the relative differences between countries in peanut sensitivity compared with egg and milk.

image

Figure 1.  Relative proportions of egg, peanut and milk allergy in European and Asian community-based populations.

Download figure to PowerPoint

Food allergen exposure in early life may not necessarily be correlated with reduced sensitization. The prevalence rate of childhood peanut allergy in Israel is low at 0.04% where children are often exposed to peanuts very early owing to the popularity of locally produced peanut snacks (25). This is in contrast to the USA and European countries such as the UK, where avoidance of peanut is encouraged until preschool age (8, 26). The incidence of food sensitization is similar among Estonian and Scandinavian infants during the first years of life; however, sensitization is more likely to persist into the later childhood of Scandinavian children suggesting differences in tolerance induction (27). Such observations suggest that we need to look elsewhere for a greater understanding of food allergen sensitization rather than to attribute it primarily to high allergen exposure in early life. The differences in food allergen sensitization in different communities may potentially result from many factors such as genetic constitution, the early life home environment, exposure to microorganisms and infant feeding and weaning practices.

Evidence from clinical trials and cohort studies

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

Clinical trials and prospective cohort studies have largely failed to demonstrate a substantial reduction in food allergy by preventing exposure to potential allergenic foods. Three areas that have received particular attention are; maternal avoidance of foods whilst being pregnant and/or whilst lactating, use of modified formula (partially, extensively hydrolysed or elemental) and delayed introduction of solids.

Although sensitization may occur via passive transfer of allergen in utero, strict allergen avoidance in maternal diet during pregnancy has been associated with either no effect, or an increased incidence of food sensitization in infants (12, 28). Furthermore, immediate hypersensitivity to peanut, milk and egg is not uncommonly found in patients upon the first known exposure, despite complete avoidance beforehand (29–32). This suggests that even strict maternal avoidance is not necessarily successful in eliminating exposure to the food allergens in question. In support of this, Vance et al. (33) found that serum levels of ovalbumin were similar in women irrespective of whether or not they were attempting to avoid strictly egg in their diet. Is it possible that the immune response to these very small amounts of allergen may be actually sensitizing rather than inducing tolerance?

Although there are some data to suggest that avoidance of egg, milk and peanut whilst breast feeding may reduce the severity of atopic dermatitis and delay the onset of food allergy, no long-term protective effect on food allergy has been established. It may be that there is a risk in advocating allergen avoidance in lactation by modifying the processes by which tolerance occurs.

A small protective effect on the development of food allergy is seen with the use of ‘hypoallergenic’ formula; however, some of these studies may be examining allergen modification rather than allergen avoidance. A recent Cochrane review (13) examining the role of formula on subsequent atopy found some evidence for a protective effect on the development of cow’s milk allergy with hypoallergenic formula in high risk infants; however, this effect was seen for both partially and extensively hydrolysed formula. Others have found a similar protective effect with use of modified formulas (34). Extensively hydrolysed formula contains tiny amounts of β-lactoglobulin (0.91 ug/l) equivalent to that in breast milk, while partially hydrolysed formula contains much greater amounts (12,400 ug/l).Thus, the use of partially hydrolysed formula is not allergen avoidance, but more allergen modification. It is not clear from primary prevention studies whether extensively hydrolysed formula is more effective than partially hydrolysed formula in prevention of cow’s milk allergy despite the greatly differing content of cow’s milk allergen (13, 35). This is also consistent with the observations in animal models where feeding a partially hydrolysed formula induced tolerance to β-lactoglobulin, while the extensively hydrolysed formula did not (36).

Delayed introduction of solid foods has commonly been promoted as a method of reducing food allergen sensitization (11). Although food allergen avoidance may delay the onset of food allergy, it does not appear to alter the prevalence in later childhood (37) and may be more successful for some allergens (e.g. milk) than others (e.g. egg) (38). In a pre- and post-natal study of food allergen avoidance, Zeiger et al. found that egg sensitization was not related to exposure in the first 2 years of life. Furthermore, the rate of acquisition of sensitivity and subsequent tolerance induction was similar in both groups despite approximately 80% of the control group and only 20% of the avoidance group being exposed to egg by 12 months of age (38). In this study, cow’s milk allergen sensitization was significantly less in those avoiding cows milk. Concerning wheat allergy, this actually increased when initial wheat exposure was delayed to over 6 months of age (39). A large population-based prospective birth cohort study failed to provide evidence of a protective effect of delayed introduction of solids past the sixth month of life on both atopic dermatitis and sensitization (RAST) to common food allergens (40).

It is possible that early and regular exposure to food allergens may decrease rather than increase the risk for subsequent food allergen and/or the natural history and tendency for resolution of that allergy. Early and regular exposure to peanut as a primary prevention measure is currently being investigated in the LEAP (Learning Early About Peanuts) study and its results are keenly anticipated (41).

Other clinical observations

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

Once tolerance is established either naturally or by oral tolerance induction with graded exposure to foods, continued allergen exposure favours the maintenance of tolerance rather than resensitization. In peanut, egg and cow’s milk allergy following oral tolerance induction or loss of sensitization, regular intake of the allergen is required to maintain tolerance and cessation of exposure can result in systemic reactions (42, 43). Furthermore, in children with egg allergy who were desensitized with graded increasing oral exposure to egg allergen, egg-specific IgE was observed to decrease over 24 months (44). Such studies do not provide support for the proposition that continued exposure to allergen will increase the IgE level or delay the acquisition of tolerance.

Avoidance of allergen in sensitized, but not clinically reactive individuals may result in an enhanced IgE response and severe allergic reactions on re-exposure. There have now been multiple case series and case reports of such occurrences. Children with severe atopic dermatitis appear to be particularly at risk for this loss of tolerance and have developed severe allergic reactions and anaphylaxis on reintroduction of previously tolerated foods (including milk, fish and peanuts) to which they were SPT positive (45, 46). Fatal anaphylaxis to inhaled cow’s milk proteins occurred in an 18-year-old girl who was sensitized, but previously tolerated cows milk (47). Similar reactions on re-exposure have been reported for atopic children with asymptomatic sensitivity to fish (48). These studies suggest that in certain circumstances, allergen avoidance may be not only unhelpful but also harmful.

Acquisition of tolerance

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

The exact mechanisms underlying acquisition of tolerance and sensitization are not understood. In the 1980s, prospective studies on allergic children demonstrated that in atopic children, tolerance to food allergens as measured by a reduction in food-specific IgE antibodies was commonly acquired after 2 or 3 yr of age, despite the ongoing allergen exposure (49, 50). In this regard, the study of Bjorksten and others is of particular importance as it demonstrated that non-atopic children also developed transient food-specific IgE responses, albeit at a lower level than atopic children (50). Furthermore, the resolution of the food-specific IgE frequently occurred in the face of continuing allergen exposure. In animal models, continued food allergen exposure has been shown to promote the development of tolerance. For example, feeding of cow’s milk protein in the form of unmodified and partially hydrolysed infant formula readily induced oral tolerance, whilst mice fed an extensively hydrolysed formula failed to develop tolerance to alpha casein (51).

T cells play a role in the development of tolerance. The balance between allergen-specific T-regulatory cells and T helper 1 and 2 cells is considered critical in the development of IgE immune responses against allergens. Children who acquired oral tolerance to peanut demonstrated Th1-skewed responses when tolerance was achieved (52). Karlsson et al. showed that children who outgrew cow’s milk allergy developed higher frequencies of circulating T-regulatory cells following a cow’s milk challenge and exhibited decreased in vitro proliferative responses to bovine β-lactoglobulin in peripheral blood mononuclear cells compared with children with persisting cow’s milk allergy (53). Children with IPEX, an inherited defect of T regulatory function develop severe and multiple IgE-mediated food allergic responses and markedly elevated IgE levels (54). The suggestion that T regulatory cells are involved in the control of childhood food allergy is of particular interest as it suggests approaches to new strategies for prevention and treatment.

An important issue to be considered in the development of tolerance or sensitization is the route of allergen presentation. Presentation of antigens via the gastrointestinal tract is more effective in inducing tolerance than via other routes. However food allergens are not only presented to the immune system via the gastrointestinal tract and epicutaneous presentation favours a Th2 type response to ovalbumin (55) and peanut (56) in animal models. Similar considerations may apply to humans. Increased levels of sensitization to grape allergens were found in workers who handled the fruit (57). The increased occurrence of peanut sensitization in children utilizing topical treatments containing peanut oil (58) and in households where peanut consumption is high implicates allergen presentation via a route other than oral ingestion (59).

Modification of allergenicity as a promoter of tolerance

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

That allergen modification may promote tolerance is demonstrated by successful studies on immunotherapy with modified allergens derived from peanut, egg and soy proteins (60). Food allergen degradation and/or in vivo modification in the gastrointestinal tract may play a role in the normal process of tolerance. It is well established that the allergenicity of foods such as egg and cow’s milk can be altered by heat and/or changes in pH. Boiling denatures the major peanut allergenic protein Arah1 to a greater extent than Arah2 and Arah3, and roasting can increase the amount and IgE binding intensities of the major allergenic peanut proteins (61). It has been suggested that preparation techniques may account for differences in peanut allergy in different countries as boiling is common in Israel and Asia, whereas roasting is common in the USA(61). However, boiling only reduces specific peanut IgE binding capacity by half and does not eliminate allergenicity (62). Thus, it is not clear whether reduced peanut sensitization observed in Asian countries, where peanuts are commonly consumed, is because peanuts are commonly boiled, as allergens are still present in boiled peanuts and amounts of total allergen consumed may be similar or even increased.

Subjects known to react to native allergen may not react after foods are heated or undergo chemical modification. Food allergy is more persistent in subjects whose IgE antibodies are directed against epitopes, which are not destroyed by heat or acid (63). Both human (64) and murine studies (65) suggest that alteration in gastric pH by antacids, H2 antagonists and proton pump inhibitors enhances the production of food-specific IgE antibodies. Decreasing gastric acidity may prevent degradation of food proteins to a less allergenic form. In light of these findings, the widespread use of proton pump inhibitors in infants for common symptoms of colic and irritability on the unproven assumption that these are frequently because of gastroesophageal reflux induced reflux oesophagitis needs further examination (66).

Evidence from animal models

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

Animal models of food allergy suggest that elimination or restriction of food allergen exposure in early life may promote IgE sensitization and conversely that larger amounts of allergen may promote tolerance. Exposure to egg albumin in rats in early life suppressed specific IgE responses (67). In already sensitized animals, large primary doses of egg antigen (greater than 100 ug) inhibited rather than promoted secondary IgE responses (68). Similar findings are observed in murine models of peanut allergy where oral sensitization is dose-dependant, with low-dose exposure to peanut protein (0.2 and 0.02 mg) leading to sensitization and high-dose exposure (100 mg) inducing tolerance (56). In a murine model of hypersensitivity to ovalbumin, repetitive oral administration of allergen induced tolerance rather than an increased level of sensitization (69).

Maternal exposure to food allergens during gestation may be beneficial by virtue of enhancing IgG responses resulting in placental transfer of the IgG antibody. Passively acquired maternal rat IgG has been shown to exert a long lasting suppressive effect on subsequent IgE sensitization in offspring (70). The suppressive effect of maternal IgG on food allergen sensitization in animals raises the possibility that avoidance of food allergens during pregnancy resulting in decreased passively acquired IgG levels may favour an increase in the food-specific IgE in the infant. It has been observed that serum IgG antibody levels against ovalbumin decreased in women who avoided egg ingestion during pregnancy (33).

T regulatory cells can reduce intestinal inflammation via transforming growth factor-beta (TGF-beta) dependent mechanisms (71) and food allergic children have reduced TGF-beta in the gut (72). In humans, it has been suggested that breast milk TGF-beta may diminish the risk of allergic disease. An allergen specific mechanism is supported by animal models. The administration of TGF-beta in conjunction with food allergen reduced OVA-specific IgE and prevented the development of anaphylaxis on challenge in mice (73) and reduced cow’s milk sensitization in rats (74). Is it possible that restriction of food allergen exposure in breastfed infants reduces allergen-specific tolerance-inducing effects of TGF-beta?

Plant compounds can alter sensitization to other food allergens. In mice, concurrent ingestion of apple polyphenols inhibits the development of oral sensitization to ovalbumin (75) and herbal extracts block peanut anaphylaxis in sensitized mice (76). Dietary simplification or restriction may also promote allergic sensitization or reactions by removing unrecognized inhibitory components.

Conclusion: Avoidance or exposure?

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References

The mechanisms which determine sensitization and tolerance in an individual are incompletely understood. The strict avoidance of food allergens for the prevention and management of sensitization has been advised for many decades. The evidence to support this advice is at the best incomplete. There are no good data to support the notion that strict avoidance in sensitized individuals of trace amounts of foods (below the threshold that provokes symptoms) delays or prevents the acquisition of tolerance.

The ideal perhaps is to develop strategies that promote tolerance. But how do we do so? We and others (77–79) suggest that a focus on complete food allergen avoidance is not necessarily correct. It is also possible that such a focus may have inadvertently contributed to an increase in food allergy. In contrast, immunotherapy strategies follow graded and regular exposure to induce tolerance followed by regular exposure for its maintenance. Exploration of other models of managing and preventing food allergy is required. Perhaps our thinking should be more directed to presentation of food allergens in a dose and form that favour the induction of natural mechanisms of tolerance. Randomized controlled studies of these issues are required to inform changes in public health measures.

In 1986, Zieger et al. concluded that ‘monumental obstacles confront preventive efforts to overcome the strength of the allergic constitution’(80). There has been an inexorable increase in food allergy in western countries since then, despite ever more stringent attempts at reduction in exposure to food allergens. We suggest that a change in perspective may be helpful.

References

  1. Top of page
  2. Abstract
  3. Historical recommendations
  4. Epidemiology
  5. Evidence from clinical trials and cohort studies
  6. Other clinical observations
  7. Acquisition of tolerance
  8. Modification of allergenicity as a promoter of tolerance
  9. Evidence from animal models
  10. Conclusion: Avoidance or exposure?
  11. References