Estimating Prevalence Of Soy Protein Allergy

Soy protein has been designated one of the major allergens by the U.S. Food and Drug Administration. Nevertheless, rigorous prevalence data are not available for soy protein allergy, and indirect evidence suggests it occurs less frequently than other common allergens including shell-fish, peanuts, tree nuts and fish.

Food allergies are becoming an increasingly important issue. True soybean allergy involves the generation of allergen-specific IgE antibodies in certain individuals upon exposure to soy protein. Soybeans contain several allergenic proteins. IgEmediated soybean allergy can be a serious condition because of low threshold doses for the offending protein(s) and the possibility of severe reactions in some soy-allergic individuals.

Soybeans are considered by worldwide regulatory authorities as one of the most commonly allergenic foods. Soybean allergy is most common among infants and young children because of exposure to soybean-based infant formula.1, 2 Many infants outgrow soybean allergy, so the prevalence is lower in adults.3

Despite widespread acceptance of the notion that soybeans are a commonly allergenic food, solid scientific information on the prevalence of soybean allergy in the overall population does not exist. However, the prevalence of soybean allergy can be estimated by extrapolation from existing information. A rigorous estimate of the prevalence of soybean allergy would need to involve several critical features: 1) a study of the general population; 2) clinical demonstration of adverse reactions to soybean preferably by double-blind, placebo-controlled food challenges (DBPCFC); 3) and clinical documentation of an IgE-mediated mechanism for the adverse reaction. The literature relevant to the prevalence of soybean allergy was screened, and no clinical studies were found that met these rigorous criteria. Thus, the prevalence of soybean allergy must be estimated from a critical examination of existing studies.

Only one study actually attempted to determine the prevalence of soybean allergy directly in the general adult population. Bjornsson et al. indicated a 2 percent prevalence of soy allergy among 1,397 unselected Swedish adults of ages 20-44 years.4 However, the diagnosis was based only on serum screening for soy-specific IgE and is thus undoubtedly an over-estimate because no oral challenges were done to confirm that adverse reactions to soy actually occurred in these individuals. It is well known that individuals can have food- specific IgE in their serum without clinical reactivity to oral challenge to that food, especially in the case of legumes, like soybean.5 In other studies, fewer than 10 percent of patients with positive soy-specific IgE levels reacted adversely on blinded oral challenge 6, 7; however these two studies cannot be used to estimate overall prevalence of soybean allergy because they were confined to food-allergic subpopulations. If we assumed that 10 percent of the Swedish subjects with soy-specific IgE would have been challenge- positive, then the prevalence of soybean allergy is estimated at about 0.2 percent in adults.

Clinical experience would suggest that the prevalence of soy allergy should be higher among infants and young children than among adults. Yet, the prevalence of soybean allergy among unselected Australian children was estimated at only 0.1 – 0.28 percent based upon a unique extrapolation strategy developed initially by Hill et al.8 They determined the prevalence of milk allergy among a cohort of high-risk infants (those born to families with allergies). Then, they corrected this estimate to the general Australian population by estimating the percentage of infants who would be considered high-risk (19.3 percent). Their estimate of the prevalence of milk allergy among Australian infants was 2 percent, which is in agreement with earlier studies of the prevalence of milk allergy in other countries based on challenge trials of the general infant population.9, 10 The prevalence of soy allergy can be estimated from the prevalence of milk allergy based upon data on the subsequent development of soy allergy in milk-allergic infants who were switched to soy formula. Various investigations estimate that 5-14 percent of milk-allergic infants develop soy allergy.1, 10, 11, 12 Thus, the prevalence of soybean allergy is estimated as 2 percent x 5-14 percent, or 0.1-0.28 percent. Information from the infant formula industry could be used to confirm these estimates because milk-allergic infants are often switched first to soybean- based infant formula. A study by Cordle2 can be used to confirm the estimate of Hill et al.8 because it provides information on a large number of milk- allergic infants switched to soybean formula. Cordle indicated that 26 out of 247 cows’ milk-allergic infants (10.5 percent) were unable to tolerate soy-based infant formula. If 10.5 percent of cows’ milk-allergic infants have soy allergy, then an estimated prevalence of soy allergy in the infant population would be 0.21 percent (2.0 percent x 10.5 percent).

Of course, this type of approach to estimating the prevalence of soybean allergy has a significant potential flaw. It assumes that only infants with cows’ milk allergy will develop soy allergy in infancy. But in all likelihood, at least for young infants (those less than 2 years of age, the age group typically exposed to infant formula), that is probably true for the vast majority of cases.

Many infants with soybean allergy will outgrow their soybean allergy within a few years. Thus, using the prevalence of soy allergy in infants to estimate the prevalence in the overall population would require a significant adjustment for this development of oral tolerance. In studies on a limited number of soy-allergic infants, 50-100 percent of the infants became tolerant of soybeans within 2-3 years.3, 13 It is thought that the development of oral tolerance to soybean follows the same pattern as for milk and egg, where much larger numbers of infants have been followed. In an evaluation of several studies, Eggleston estimated that 70 percent of children who developed food allergies at ages less than 3 years would outgrow the allergy within a few years.14 Most of these infants were sensitive to milk and egg, although some were allergic to soy or wheat. In a preceding paragraph, the estimated range of soybean allergy in infants was 0.1-0.28 percent. If we assume that 70 percent of these infants will outgrow their soybean allergy, the estimated prevalence in the overall population would be 0.02-0.056 percent. Of course, this assumes that all sensitization to soybeans is initiated in infancy. While this is not true, the majority of soy allergy likely does emanate from sensitization during infancy.

Since most soy allergy is outgrown, the prevalence of soybean allergy should be higher among infants than among adults. The only study of the prevalence of soybean allergy among adults is the estimate of 0.2 percent for Swedish adults – higher than predicted based on this logic.

Despite the continuing uncertainties, the prevalence of soybean allergy in the general population is probably not higher than 0.2 percent and could be as low as 0.1 percent (290-580,000 individuals) among the U.S. population. This likely prevalence for soybean allergy contrasts with prevalence estimates of 1.9 percent for crustacean allergy, 0.6 percent for peanut allergy, 0.5 percent for tree nut allergy, and 0.4 percent for fish allergy.15, 16 Editor’s Note: This work is a contribution from the University of Nebraska Cooperative Extension Division, Lincoln, Nebraska, Journal Series No.1034.



Steve L. Taylor, Ph.D. is a professor in the Department of Food Science and Technology at the University of Nebraska, and is co-director of the Food Allergy Research and Resource Program. He received his doctorate in biochemistry from the University of California – Davis.


1) Zeiger RRSS,, SSaammppssoonn HHAA,, BBoocckk SSAA,, eett aall.. JJ.. Pediatr.1999; 134:614-622. 1J. Pediatr 999; .
2) Cordle CT. J. Nutr. 2004; 134:1213S-1219S.
3) Sampson HA, Scanlon SM. J. Pediatr.1J. Pediatr 989;115:23-27. . 1989;
4) BjornssonE,JansonC,PlaschkeP,etal.Ann. AllergyAsthmaImmunol.1996; 77Bjor :327-332. nsson E, Janson C, Plaschke P, et al. 77Ann. Allergy Asthma Immunol. 77:
5) Bernhisel-Broadbent J, Sampson HA. J. AllergyClin.Immunol. 1989; 8J. Aller 3:435-440. gy Clin. 83:
6) GiampietroPG,RagnoV,DanieleS,etal.Ann. Allergy1Giampietr 992; 69:143-146. o PG, Ragno V, Daniele S, et al. Allergy1Ann. Aller gy 1992;
7) Magnolfi CF, Zani G, LacavaL,etal.Ann. AllergyAsthmaImmunol.1996; 77Magnolfi CF :197-201. , Lacava L, et al. Allergy Asthma Immunol. 77:
8) Hill DJ, Hosking CS, Zhie CY, et al. Environ. Toxicol. Pharmacol. 1997;4:101-110.
9) Jakobsson I, Lindberg T. Acta Paediatr. Scand. 1979; 68:853-859.
10) Host A, Halken S. Allergy 1990; 45:587-596.
11) Cantani A, Ferrara M, Rango V, Businco L. Riv. Eur. Sci. Med. Farmacol.1990;12:311-318.
12) Klemola T, Vanto T, Juntunen-Backman K, et al. J. Pediatr.2002J. Pediatr ; 140:219-224. . 2002;
13) Bock SA. J. AllergyClin.Immunol. 1Bock Aller 982; 69:173-177. gy Clin. 1Immunol. 1982;
14) Eggleston PA.Ann. Allergy1987; 59(5 Pt. 2):1Eggleston P 79-182. A. Allergy 1987; 179-
15) SichererSH, Munoz-Furlong A, BurksAW,SampsonHA.J. AllergyClin.Immunol. 1Sicher 999; 103:559-562. er SH, Burks AW, Sampson HA. Allergy Clin. 1999;
16) SichererSH, Munoz-Furlong A., SampsonHA.J. AllergyClin.Immunol. 2004Sicher ; 114:159-165.
15) SichererSH, Munoz-Furlong A, BurksAW,SampsonHA.J. AllergyClin.Immunol. 1Sicher 999; 103:559-562.

Facebook comments