Article written: March 2002
Cross-reactivity occurs when the immune system mistakes a similar protein or chemical composition for an allergen, causing an adverse reaction. The immune system may react to foods or plants in the same botanical family. For example, a person who cannot tolerate peach may not be able to eat apricots; or a person allergic to birch will often get oral symptoms from eating fresh apples.
Clinical implications of cross-reactive food allergens
(Adapted from Current reviews of Allergy & Clinical Immunology by Scott H Sicherer, MD, JACI, Dec. 2001)
The Allergist is called on to determine the risk of reaction to related foods among legumes, tree nuts, fish, shellfish, cereal grains, mammalian and avian food products, and a variety of other plant-derived foods that may share proteins with pollens, latex, and each other. Clinical evaluation requires a careful history, skin prick tests and RAST-type blood tests, and in some cases oral food challenges. The pitfalls in evaluation are interpreting the clinical relevance of a positive skin prick test or RAST and deciding when it is a false-positive result.
Cross-Reactions Among Various Foods
It is common to find positive responses for IgE to several beans in individuals who are clinically reactive to one type. Using RASTs, Barnet et al screened sera from 40 patients with peanut allergy against 10 other legumes and demonstrated IgE binding to multiple legumes for 38% of patients. Bernhisel-Broadbent and Sampson studied 62 children with allergy to at least 1 legume and found that 79% had serologic (positive RAST) of IgE binding to more than 1 legume, and 37% binding all 6 legumes.
Despite the high rate of cross-sensitization, clinical cross-reactions are uncommon, as demonstrated by studies of allergenic legumes, such as peanut and soy. Among 113 children with atopic dermatitis evaluated with DBPCFCs only 1 (0.8%) had clinical allergy to both foods, despite 19% reacting to peanut and 5% to soy. Bock and Atkins studied 32 children with peanut allergy confirmed by DBPCFC and found that 10 (32%) had a positive skin test response to soy, but only 1 (3% of those with peanut allergy) had a clinical reaction.
In considering a wide variety of legumes, only 3 (1.8%) of 165 children with atopic dermatitis evaluated with DBPCFCs reacted to more than 1 legume, despite 19% reacting to at least 1 legume. Bernhisel-Broadbent and Sampson specifically addressed the issue of legume cross-reactivity by performing open tests or DBPCFCs in 69 highly atopic children with at least 1 positive skin response to a legume. Oral challenge to the 5 legumes (peanut, soybean, pea, lima bean, and green bean) resulted in 43 reactions in 41 patients (59%). Only 2 (5%) of 41 with any 1 positive challenge reacted to more than 1 legume. The authors concluded that elimination of all legumes in individuals with clinical reaction to 1 legume was unwarranted, despite the high prevalence of patients with multiple legume-positive skin prick tests (SPT) responses.
These studies did not include large batteries of legumes, and it may be that particular types are more allergenic or cross-reactive. In an evaluation of children with peanut allergy in France, 11 (44%) of 24 had positive skin prick test response to lupine, and of 8 subjects who underwent DBPCFC (6 children) or labial challenge (2 children) to lupine, 7 reacted.
Regional dietary habits and pollen exposure may influence the epidemiology of legume allergy. In Spain, for example, allergy to lentil was more common than allergy to peanut, and of 22 children with lentil allergy evaluated for reactions to other legumes, 6 had a history of reacting to chickpea, 2 to pea, and 1 to green bean. These findings raise suspicion for multiple legume allergy on those reacting to lentil, lupine, and chickpea, but more studies in a variety of geographic settings are needed to quantify the risk.
Assessment of cross-reactivity among tree nuts is complicated by shared allergens among the nuts and between nuts and other plant-derived foods and pollens. Clinical reactions to tree nuts can be severe, potentially severe and can occur from a first exposure to a nut in patients allergic to other nuts. Studies have indicated a high degree of IgE binding to multiple tree nuts.
In the author’s study of children with tree nut allergy, 92% of 111 patients with peanut allergy, tree nut allergy, or both had IgE antibody to more than 1 tree nut, and 37% of 54 had experienced convincing reactions and had specific IgE antibody to more than 1 nut.
Bock and Atkins performed challenges to 1 or more nuts in 14 children, and at least 2 reacted to multiple nuts (as many as 5 types). Several studies have reported coallergy to multiple tree nuts. In Ewan’s study coallergy was seen in over a third of 34 patients evaluated for tree nut allergy. Considering the potential severity of the allergy and issues with accurate identification of particular nuts in prepared foods, caution would seem sensible, and total elimination of the nut family is suggested. Several clinicians (including the author) suggest that if a nut, eaten in isolation, is previously tolerated by a nut- allergic individual it would be OK to continue eating this nut. Other clinician’s (including myself) take a more restrictive approach, to avoid accidents and also, because there are reports of nut allergic individuals becoming allergic to nuts that they previously tolerated. It is known that some nuts are homologous and cause reactions (eg, in pistachio-cashew), whereas others may be homologous but rarely elicit clinical cross-reactivity (eg, proteins in coconut and walnut).
Legumes, tree nuts and seeds
Cosensitization to allergenic foods, such as peanut, tree nuts, and seeds (sesame, poppy & mustard) is common. In a study of 731 subjects in the UK, 59% sensitised to peanut were sensitised to hazelnut, Brazil nut, or both. Although clinically significant cross-reacting proteins have not yet been described, coallergy to peanut and tree nut has been reported between 23% and 50% in referral populations of atopic patients. The rate of coallergy is much lower in unselected populations (2.5%). The clinician must consider the age of the patient, history, and perhaps sensitisation in considering categoric elimination of these allergenic foods. Reactions to seeds, such as sesame, mustard, and poppy, are reported, and cross-reactivity with foods (hazel, kiwi, and other seeds) and pollens is potentially important, but the full clinical implications are far from established.
Several reports demonstrate that isolated allergy to a single species of fish (eg, tropical sole and swordfish) occurs and usually does so in the relative absence of IgE antibody to common fish allergens (Gad c 1). However, positive skin test responses to multiple fish in subjects with fish allergy is almost the rule, and clinical cross-reactivity is also common. In 61 children with a history of fish allergy exposed to 2 to 8 species, 34 (56%) reacted to all, and 27 (44%) tolerated some types.
Summary: A patient with fish allergy is at high risk for reactions to other fish but may tolerate some fish species and may deserve further evaluation with supervised oral challenges if desirous of ingesting other fish. The fact that fish allergy can be severe and that cooking-canning and other processing can alter allergenicity must be considered during these evaluations.
Invertebrate tropomyosin is a panallergen with significant sequence homology identified in Crustacea, such as shrimp, crab, and lobster; molluscs, such as oyster, scallop, and squid; parasites, such as Anisakis and insects, such as cockroach, grasshopper, and dust mite, with less homology to vertebrate tropomyosin. Although the clinical impression is that reactions to multiple crustaceans are fairly common, there are few clinical studies addressing this issue. In 16 atopic patients with shrimp allergy, greater than 80% had positive SPT responses to crab, crayfish, and lobster. In 11 patients with immediate reactions to shrimp ingestion, the reaction rate to crab, lobster, and crayfish was 50% to 100% per species. On the other end of the spectrum is a report of several individuals with reactions to only particular species of shrimp. Overall, Crustacea represent an increased risk of cross-reactivity, with a potential for severe reactions.
Even les well defined is the risk for mollusk allergy for individuals with allergy to Crustaceae or mollusk. Even though the studies are not conclusive, the overall impression is that the risk of mollusk cross-reactivity is common in patients allergic to Crustacea.
Tropomyosin is found in house dust mite, an aeroallergen, which raises the possibility of sensitisation by the respiratory route. In a report of asthma induced by snail consumption in 28 patients, RAST inhibition studies indicated that house dust mite sensitisation was the likely sensitising event. There are several reports linking immunotherapy (IT) with house dust mites to the development of severe reactions to molluscs and crustacea.
Cereal grains (eg, wheat, barley, rye, and oat) share homologous proteins with grass pollens and each other. This may account for the high rate of cosensitization to these foods, but among 145 children with positive Skin prick test responses to cereal grains, only 21% exhibited clinical reactivity during challenges. In addition, among those with reactions to 1 grain, 80% were tolerant of all other grains. Caution is warranted, but clinical reactivity to multiple grains appears uncommon.
Mammalian and avian food products
Cross-sensitization is more common within than between mammalian meats, but clinical correlation with sensitisation is generally under 50%. For avian foods, sensitisation has been described to a-livetin found in feathers, egg and meat, and associated with reaction to chicken meat in 22% to 32%. When chicken meat allergy is present without egg allergy, the risk of reaction to multiple species of avian meats (turkey, pheasant, and quail) may be increased.
A study with oral challenges showed that 9.7% of 62 children with cow’s milk allergy (CMA) reacted to beef. Heating reduces the allergenicity of beef, and therefore, well-cooked beef is less likely to cause a problem with those with CMA.
In vitro studies shows extensive cross-reactivity among sheep’s, cow’s, and goat’s milk and among cow’s, ewe’s, goat’s, and buffalo’s milk, with no significant binding to camel’s milk. 92% of 26 patients with cow’s milk allergy reacted to goat’s milk during oral challenge. However, only 4% of 25 children with cow’s milk allergy rteacted to mare’s milk.
Oral allergy syndrome (OAS) is classically described as isolated oral symptoms caused by labile proteins in fresh fruits and vegetables that share homology with proteins in pollens (the initial source of sensitisation). Several clinical associations have been described (eg, birch pollen with Rosaceae fruits, ragweed with melon, and mugwort with celery). The number of foods reported to be involved in the syndrome is ever increasing. Cooked forms of the foods are typically tolerated. The epidemiology varies by the exposure to pollens. 23% to 76% of patients with allergic rhinitis experience OAS. More than70% of patients with OAS react to more than 2 foods. Peach is the dominant allergenic fruit.
In a review of several studies with a total of 1,361 patients allergic to food pollen with OAS, 8.7% experienced associated systemic symptoms outside of the gastrointestinal tract, 3% at some time experienced systemic symptoms without oral symptoms, and 1.7% experienced anaphylactic shock. Hence the term pollen-food syndrome may be more appropriate than OAS. There is evidence that when fruit allergy develops in the absence of pollen allergy, reactions are directed tolipid transfer proteins (LTPs). Reactions involving fruits with homologous LTPs are more likely to be severe. One study compared patients with Rosaceae fruit allergy with and without pollinosis (hay fever) and found that systemic reactions occurred in 82% without compared with 45% with pollinosis. Anaphylactic shock was also more common in the former (36% vs 9%, respectively). A similar theme was noted for hazelnut, in which patients without pollinosis experienced severe reactions and had IgE binding to hazelnut that were heat stable. One study found that individuals with positive skin prick test responses to commercial Rosaceae food extracts (presumably enriched for stable allergens) were more likely to experience systemic reactions than those with responses positive only to fresh extracts.
Foods commonly reprted to cross-react with latex include banana, avacado, kiwi, chestnut, potato, and papaya. In a study of 136 patients with latex allergy evaluated by means of RAST to 12 foods reported to be involved in latex-food reactions, 69% of responses were positive to at least 1 food, and 49% were positive to more than 1 food. In one study of 47 patients with latex allergy, 100 of 376 food skin test responses were positive, but only 27 (7.2%) were associated with clinical reactions. In patients with isolated latex allergy, reactions are more likely to banana, avacado and kiwi, whereas those with latex allergy & pollinosis are more likely to react to Rosaceae foods and celery.