Soy Protein Intolerance

Updated: Oct 06, 2017
Author: Stefano Guandalini, MD, AGAF; Chief Editor: Carmen Cuffari, MD 



Soy-based formulas were introduced in infant nutrition 100 years ago, when their use was recommended for the treatment of summer diarrhea. Eighty years ago, the use of soy-based formulas was extended to the treatment of cow's milk intolerance. In the 1970s, use of soy-based formulas became common; in the 1970s and 1980s, US consumption of soy-based formulas was around 25% of that of cow's milk–based formulas.

In the last few years, interest in soybeans and soybean components has markedly increased, mainly because of the potential influence of soy on the development of heart disease, cancer, kidney disease, osteoporosis, and menopause symptoms. Unfortunately, soy protein formulas (SPFs) can cause allergies and other intolerance reactions. For many years after the first description by Duke in 1934, soy was considered a weak sensitizing protein based on animal study findings. In the 1960s, several other authors confirmed the potential allergenicity of soy protein formulas.

A higher prevalence of soy intolerance has generally been reported in non–immunoglobulin E (IgE)-associated enterocolitis and enteropathy syndromes. Authorities have failed to reach consensus on the risk of feeding allergic or nonallergic infants with soy protein milks.[1, 2] This divisive clash of opinion is also reflected in the mutually antagonistic stances adopted by 2 important scientific societies, the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) and the European Society of Pediatric Allergy and Clinical Immunology (ESPACI).

However, the general agreement is that a significant number of children with cow's milk protein intolerance develop soy protein intolerance when soy milk is used in dietary management. For this reason, the American Academy of Pediatrics (AAP) and ESPGHAN recommend the use of extensively hydrolyzed or free amino acid-based formulae in the treatment of cow's milk protein allergy.[3] According to ESPGHAN, soy protein formula should particularly not be used in infants with food allergy during the first 6 months of life.[4]

However, the AAP states that infants with IgE-associated symptoms of cow's milk allergy may benefit from a soy formula because the risk of cross-reactivity does not appear to be very high.[3] A Cochrane systematic review confirms that soy formula cannot be recommended for prevention of allergy or food intolerance in infants.[5]


Two heat-stable globulins constitute 90% of the pulp-derived proteins: beta-conglycinin, which has a molecular weight (MW) of 180,000, and glycinin, which has an MW of 320,000. Immunoblotting and competitive enzyme-linked immunosorbent assays have identified a 30-kD glycinin from soybeans that cross-reacts with cow's milk caseins and is composed of 2 polypeptides (A5 and B3) linked by a disulphide bond. The protein's capacity to bind to the different antibodies relies on the B3 polypeptide.

However, other soy proteins can act as allergens in humans. At least 9 proteins with MW ranging from 14,875-54,500 were found to react with human IgE in patients with asthma. Moreover, after enteric digestion, numerous potential antigens are generated at the mucosal surface.

According to some animal study findings, soy proteins appear to be less sensitizing than cow's milk proteins; however, infants with a previous history of cow's milk protein intolerance have a greater risk of developing soy protein intolerance. The intestinal mucosa damaged by cow's milk proteins may allow increased uptake of the potentially allergenic soy proteins.

Antigenicity of soy-based products is strongly influenced by methods of preparation; therefore, clinical manifestations can be elicited by some soy-based products and not others.


All soybean proteins and foods currently available for human consumption contain significant amounts of the isoflavones daidzein and genistein, either in the unconjugate form or as different types of glycoside conjugates.

The isoflavones have structural homology to steroidal estrogens; therefore, they are considered to be phytoestrogens, but little is known about their biological activity. Unquestionably, isoflavone ingestion can elicit biological effects; however, isoflavones and their metabolites have biological properties that are quite separate from classic estrogen action.

Genistein is a potent inhibitor of tyrosine kinases and can interfere with signal transduction pathways. The threshold intake of dietary estrogens necessary to achieve a biological effect in healthy adults appears to be 30-50 mg/d.

In soy flours and concentrates, isoflavone concentrations are relatively high (0.5-3 mg/g). In soy milk and soy-based infant formulas, the concentration of isoflavones is lower (0.3-0.5 mg/g) but is 10,000-fold higher than the concentration found in breast milk. Moreover, the volume intake of these products is sufficient to account for a significantly high dietary intake of isoflavones. Infants fed soy-based formulas have plasma concentrations of isoflavones that are 3000- to 22,000-fold higher than plasma concentrations of estradiol.

Even if these substances have a weak estrogenic activity compared with estradiol, they could have adverse effects; however, the concerns about the adverse role of phytoestrogens in the first months of life are exclusively theoretical. At this time, the very limited available evidence from adult and infant populations indicates that dietary isoflavones in soy-based infant formulas do not adversely affect human growth, development, or reproduction.

The results of a study that enrolled 48 children (mean age, 37 mo; range, 7-96 mo) suggest that long-term feeding with soy protein formulas in early life does not produce estrogenlike hormonal effects.[6] No developmental problems were observed in a cohort of 129 soy protein–based formula–fed infants.[7] However, according to the Center for the Evaluation of Risks to Human Reproduction (CERHR), the possibility that adverse effects might occur cannot be dismissed. Without conclusive findings in humans, ESPGHAN recommends reducing the content of phytoestrogens in soy protein formulas because of uncertainties regarding safety in infants and young children.[4]



United States

In a national survey of pediatric allergists, the prevalence rate of soy protein allergy was reported to be 1.1%, compared with a 3.4% prevalence rate of cow's milk protein allergy.[8]


In a prospective study of healthy infants fed soy-based formula, allergic responses to soy were documented in 0.5% of infants.[9]

In a group of 243 children who were born to atopic parents and who received soy protein formula for the first 6 months of life to prevent cow's milk allergy, 14 (6%) of the children had positive skin test prick reactions to soy.[10] Only 1 of these 14 children reacted to the double-blind placebo-controlled oral food challenge to soy.

The prevalence of food allergy in patients with atopic dermatitis varies with age and the severity of atopic dermatitis. Different prevalence rates have been reported; however, in most series, 30-40% of the patients received a diagnosis of food allergy. In a study from Italy, a positive radioallergosorbent assay test (RAST) result to soy was found in 25% of children with atopic dermatitis, but a positive challenge test result to soy was elicited in only 3% of the patients.[6] Two other studies documented soy positivity in 5% of 204 patients[11] and in 4% of 143 children.[6] See the image below.

Typical atopic dermatitis on the face of an infant Typical atopic dermatitis on the face of an infant.

In a group of 93 children with documented IgE-associated cow's milk allergy who received soy formula, 14% developed soy allergy.[12] Among 35 children with food-protein enterocolitis syndrome diagnosed in a single center of Australia during a 16-year period, 34% had soy protein intolerance.[13]

In 1990, one of the authors reviewed the evidence obtained from 2108 Italian children with proven cow's milk protein intolerance and non–IgE-associated enterocolitis and enteropathy syndrome.[14] Forty-seven percent of the patients had to discontinue soy formulas because of intolerance. A higher prevalence was noted in infants younger than 3 months (53%). Thirty-five percent of children older than 1 year developed soy intolerance.

A soy-based formula is often substituted for cow's milk in infants recovering from acute gastroenteritis; however, in a previous study that recruited 18 infants with acute gastroenteritis, 3 (16%) of the children developed a clinical reaction to soy challenge, and 7 (38%) of the children developed histologic and enzymologic changes after soy challenge.[15]


Anaphylactic reactions to soy proteins are extremely rare; however, a population study in Sweden from 1993-1996 reported 4 deaths caused by soy.[16]


The risk of developing soy protein intolerance decreases with age. Among children with cow's milk protein intolerance, infants younger than 3 months are at higher risk for developing soy protein intolerance (53%) compared with children older than 1 year (35%).




The typical presentation is that of an infant who develops atopic dermatitis or cow's milk protein intolerance, which resolves with substitution of a soy-based formula but recurs 1 or 2 weeks later. Parents may report a recrudescence of dermatitis or GI symptoms. Usually, the infant presents with watery diarrhea and vomiting.

Soy protein intolerance may cause different clinical syndromes, both immunoglobulin E (IgE)-mediated and non–IgE-mediated, ranging from skin, GI, or respiratory tract reactions up to anaphylaxis. These reactions include the following:

  • Rhinitis

  • Urticaria or angioedema

  • Asthma

  • Atopic dermatitis

  • Food protein–induced enterocolitis syndrome (FPIES)

  • Intestinal villous atrophy (malabsorption syndrome)

  • Eosinophilic gastroenteritis

  • Allergic proctocolitis

  • Constipation

  • Anaphylaxis (rare)

A European multicenter study, which included both children and adults (mean age, 26.4 y; range 1-69 y), showed that bronchial asthma and seasonal rhinoconjunctivitis are the most frequent symptoms (65% of patients), followed by atopic dermatitis (33%). In this group of patients, the first reaction to soy occurred at a mean age of 19 (±10) years (range, 3-44 y).[17]

In young children and in infants, soy protein intolerance occurs mainly with dermatological and GI manifestations. Some children present with atopic dermatitis as a major symptom; however, most patients present with profuse vomiting and watery diarrhea.

Soy proteins can cause GI manifestations similar to those described in the Medscape Reference article Protein Intolerance. Although the prevalence of soy protein allergy has been traditionally considered to be quite high (for cross-reactivity) in infants and children with milk protein allergy presenting with FPIES, a large epidemiological investigation failed to confirm such high prevalence.[18]

GI symptoms are very frequent in children with cow’s milk proctocolitis or enterocolitis who are fed with soy proteins. Symptoms usually begin within 2 weeks of the infant's first feeding with soy-derived milk. Sometimes mucus can be present in the stools, but blood is rarely noted. Even if frank manifestations of colitis are absent, inflammatory changes in the colonic mucosa are frequently encountered.

Small-bowel atrophy has been documented in different studies. The degree of villous atrophy may be similar to that found in celiac disease. The mucosal damage may lead to malabsorption, hypoalbuminemia, and failure to thrive.

Some breast-fed infants can present with red blood mixed in stools as a result of soy allergic proctocolitis. These infants usually appear healthy, and hematochezia is the only symptom. In some cases, the syndrome, which can be traced down to maternally ingested soy in about 30% of cases, does not respond to withdrawal of food allergens from the maternal diet.[19]

A case of recurrent intussusception in an infant with a patch test positive to soy has been reported. The intussusceptions, which were considered a consequence of lymphoid hyperplasia, resolved with die and recurred after reintroduction of soy.[20]

Food allergies (including to soy protein) may lead to constipation. Removal of the allergen from the diet is typically very effective.[21]

The possibility of anaphylaxis occurring in adults ingesting generic drugs containing soybean oil has also been reported.[22]

In children, approximately 50% of those with soy allergy can outgrow their allergy by age 7 years. Absolute soy IgE levels were useful predictors of outgrowing soy allergy.[23]


The physical examination findings depend on the clinical picture and the duration of symptoms.

  • The most frequent presentation is enterocolitis syndrome; therefore, the infant appears dehydrated, with weight loss and sunken eyes.

  • In case of proctocolitis, the infant usually appears healthy and has normal weight gain.

  • In the less frequent case of soy-induced enteropathy, the infant has a low weight-to-length ratio and usually presents with dystrophia.

  • The signs and symptoms are related to the degree of the malnutrition. For example, edema is related to hypoalbuminemia, dermatitis enteropathica is related to low zinc level, and rickets is related to vitamin D deficiency.


According to some studies in animal models, soy proteins appear to be less sensitizing than cow's milk proteins. However, because a 30-kD protein in soy may induce cross-reactivity to cow’s milk caseins, infants with a previous history of cow's milk protein intolerance have a greater risk of developing soy protein intolerance.

The intestinal mucosa damaged by cow's milk proteins may allow increased uptake of the potentially allergenic soy proteins.



Diagnostic Considerations

Consider the following:

Differential Diagnoses



Other Tests

Soy-induced GI symptoms are not usually immunoglobulin E (IgE)-mediated; therefore, both skin tests and determination of specific IgE in serum have a low diagnostic value.

Radioallergosorbent assay test (RAST) appears to be of poor predictive value. Many children with positive results do not react to challenge tests.

Prick tests have little predictive value. The acidic subunits of glycinin and beta-conglycinin appear to be present in reduced amounts or absent in some commercial soybean skin test extracts tested by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting. As a consequence, these commercial extracts are less sensitive than extracts of soy flour.

Patch testing may provide more clinical relevant informations, particularly in children with eczema.

The role of atopy patch test in children with food allergies (including to soy) presenting with constipation has also been suggested.[21]

The challenge test with soy proteins, after an elimination diet, is the only reliable method of evaluating soy protein intolerance.



During the workup for differential diagnoses, upper or lower GI endoscopies are often performed in patients with soy protein intolerance. However, findings are nonspecific, most commonly minimal, and, at times, even completely unremarkable. Accordingly, and because of the transient nature of the disorder, endoscopies are not considered essential.


Macroscopically, only minimal erythematous changes may be observed.

Microscopically, any area (eg, lower esophagus, gastric body, antrum, duodenum) may or may not show signs of acute inflammation.

In a minority of patients, an infiltrate of eosinophils is observed.

When the clinical presentation is that of a malabsorption syndrome, the duodenal mucosa may have changes (eg, partial villous atrophy, crypt hyperplasia) indistinguishable from those of celiac disease.


Macroscopically, changes may vary from minimal erythematous segments, most commonly diffusely involving the distal colon, to severe inflammation with bleeding ulcers and loss of vascular markings.

Microscopically, nonspecific acute inflammatory changes are observed, typically indistinguishable from infectious colitis. Rarely, eosinophils predominate in the lamina propria.



Medical Care

Children affected by soy protein intolerance respond quickly to elimination of soy formula and introduction of a hydrolyzed protein formula.

Numerous therapeutic strategies have become to be investigated targeting foods that most frequently provoke severe IgE-mediated anaphylactic reactions (peanut, tree nuts, and shellfish) or are most common in children, such as cow's milk, soy, and egg. Approaches being pursued are both food allergen–specific and nonspecific. The former include oral, sublingual, and epicutaneous immunotherapy (desensitization).[24]



Medication Summary

Drug therapy is not currently a component of the standard of care for soy protein intolerance.




Soy protein intolerance is similar to other food protein intolerances. Its risk peaks during infancy, and it usually regresses completely during the first 2-3 years of life. Most children, therefore, can resume consumption of soy proteins by age 5-7 years.[23]

Patient Education

Use of soy protein formula (SPF) during the first 3 months of life does not reduce the frequency of cow's milk intolerance after the introduction of cow's milk formula.

Routine use of soy protein formula has no proven value in the prevention of atopic disease.

Routine use of soy protein formula has no proven value in the prevention or management of infantile colic.