Potential Risks

Soy protein is a safe and nutritious substance when consumed in amounts two to three times the effective daily intake proposed to achieve a cholesterol-lowering response (Goldberg et al, 1982)(2). However, increased consumption of any substance could potentially be associated with adverse effects in some individuals which might not be observed at low or moderate intakes. The potential adverse effects associated with ingestion of large amounts of soy protein which have been identified include allergenicity, fluctuations in reproductive hormones, decreased protein, decreased mineral bioavailability, and exposure to trypsin inhibitors. However, the data do not support that any one of these would pose a substantial threat to the health of the U.S. population.

Allergenicity. As a foreign protein entering the body through the gastrointestinal tract, soy protein has a potential for eliciting an allergic reaction. The 2-S globulin fraction of glycinin and beta conglycinin are believed to be the allergens responsible for hypersensitivity reactions which develop in some individuals exposed to soy protein (Leiner, 1981). Use of heat or hot aqueous alcohol in the processing of soybeans destroys the immunochemical reactivity of most protein components. Despite these precautions, a small percentage of infants dependent on soy formulas may experience adverse reactions to soy protein as an allergen largely because they have an immature immune system (Van Sickel et al, 1985).

Soy foods, in addition to cow’s milk, wheat, peanut, egg, and fish are not recommended for children younger than two years who have been identified to be at high risk for developing food allergies. High risk children include those who have exhibited atopic diseases such as asthma, rhinitis, and eczema or who have a parent who has a history of these diseases (Sampson and McCaskill, 1985). Most children eventually outgrow their food allergies over time (Sampson and Scanlon, 1989). Soy and seafood allergies are among those likely to be outgrown in contrast to allergies to milk, egg white, or peanuts.

Food allergies typically develop during infancy or in young children. Onset in children at ages older than four or in adults is an unusual occurrence because sensitization to allergens is considered a manifestation of an immature digestive tract, unless the individual has a family history and no previous exposure to the allergen (Buscinco et al, 1993). It is not known whether the incidence of soy protein allergies will increase if the numbers of individuals currently consuming this protein source increase. However, such a possibility seems remote since even among infants, the population group at highest risk of developing allergies of any kind, the prevalence of soy protein allergies is very small.

Hormonal Disturbances. As has been demonstrated by the research presented in this petition, soy isoflavones found in the protein fraction contribute to the cholesterol-lowering effects observed with ingestion of soy protein. Concern that these weakly estrogenic-antiestrogenic isoflavones may influence hormone levels in humans was initially raised more than 50 years ago by a report of infertility in sheep grazing on subterranean clover in western Australia (Moule et al, 1970). This particular clover (Trifolium subterranean L.) is a source of isoflavones. However, research on the reproductive effects of soy isoflavones which assessed reproductive hormone concentrations and organ weights at necropsy in primates did not reveal any adverse findings related to reproductive ability in either males or females (Anthony et al, 1996b, Honore et al, 1997). Other data from primates have also indicated that the estrogenic effects of soy isoflavones may be selective, affecting breast tissue, but not reproductive tissue, in surgically postmenopausal macaques (Cline et al, 1996). Further, most domesticated animals and fowl are fed soy-based chow rations and fertility is not a reported problem.

Limited data in humans suggest that ingestion of soy isoflavones may actually be beneficial for adult women. Soy phytoestrogens are believed to be the protective factor responsible for the low rates of breast cancer among women in populations where large amounts of soy protein are regularly consumed (Lee et al, 1991). In one controlled study (Cassidy et al, 1994), the effect of ingesting 45 mg of isoflavones daily from 60 g of soy protein for a period of one month was examined in six nonvegetarian premenopausal women between the ages of 21 and 29. The changes in menstrual cycle length and hormone levels observed in these women were similar to those reported in response to treatment with tamoxifen, which is currently being tested as a prophylactic agent for breast cancer.

Protein Quality. Substitution of soy protein for a proportion of the animal protein content of the U.S. diet will change the amino acid composition of the diet. Because lower biological values have been attributed to plant protein, this change may be a cause of concern for some individuals. The validity of these concerns can be challenged on face value by the more optimal health status found among vegetarians compared with the general U.S. population (White and Frank, 1994). Furthermore, the biological value of soy protein is superior to other plant proteins and is equivalent to animal protein sources (FAQ/WHO, 1991). The quantities and proportions of essential amino acids provided by soy protein are sufficient to meet human needs from age 2 to adulthood.

The previously-held belief that soy protein had a lower biological value than animal protein was based on analytical data demonstrating that methionine was a limiting amino acid in soy protein. However, these data were derived from older methods of assessing protein quality that are not the standards used today. Prior to 1993, protein quality was evaluated by calculating a protein efficiency ratio (PER) which measured the growth response of weanling rats fed different levels of a protein. This index substantially underestimates the quality of soy protein because the requirement for sulfur-containing amino acids is much higher for rats than for humans. Rapidly growing rats need greater amounts of methionine than do humans to support growth of body hair (Steinke and Hopkins, 1983). Consequently, it is estimated that the amount of methionine needed to meet human growth requirements are as much as 50% lower than the requirement for animals.

Since 1993, protein quality has been evaluated by use of the protein digestibility-corrected amino acid score (PDCAAS), which was adopted by FDA to replace the PER for food labeling purposes. The PDCAAS is recognized by the Food and Agriculture Organization and the World Health Organization as a more accurate standard for assessing protein quality than the previously used PER. The PDCAAS takes into account protein digestibility, amino acid profile, and the ability of the amino acid profile to meet the needs of 2-5 year old children, the population subgroup having the highest protein needs other than infants. Using the PDCAAS, the protein quality of isolated soy protein is identical to that of casein and egg white, and higher than that of proteins found in beef, kidney beans, pinto beans, lentils, peanuts, and wheat (FAOIWHO, 1991).

Decreased Mineral Bioavailability. Soybeans are rich sources of phytic acid and dietary fiber both of which have well-documented effects on reducing bioavailability of divalent minerals. The extraction of soy protein from soybean flakes to produce ISP and other sources of soy protein retains some part of the phytate and fiber components with the protein fraction in amounts thatvary with the processing method used. Lectins, another component of soybeans, were believed to interfere with nutrient absorption by binding to the intestinal wall (Leiner 1979). However, lectins have not been found to adversely affect growth, which indicates that any effects they may have on nutrient availability, are not biologically significant (Leiner, 1981).

In general, bioavailability of minerals from plant sources is typically lower than from animal sources. Absorption of divalent minerals such as calcium, magnesium, zinc, copper, and iron appears to be less efficient when consumed from leguminous plants such as soybeans, but ingestion of minerals from other dietary sources concurrently with soy protein does not reduce the bioavailability of minerals provided by these sources (Leiner, 1981). This observation is an important one because it supports the practice of mineral fortification of products containing soy protein to compensate for reduced availability of minerals naturally provided in these soy foods (Erdman, 1981).

Most of the data regarding the effects of soy protein on mineral balance have been obtained from studies using animal models. Studies in human subjects have yielded less conclusive results and suggest that the effects of soy protein ingestion on mineral balance in humans can not be predicted from animal studies (Erdman, 1981). The research to date suggest that zinc and iron nutriture may be most significantly affected by ingestion of soy products (Erdman and Fordyce, 1989). The primary component responsible for this reduced bioavailability is phytate which has strong mineral-chelating properties (Erdman, 1981; Erdman and Fordyce, 1989). Phytate is distributed throughout the intact soybean, but when the soybean cotyledon cell is disrupted during processing, it complexes with the primary soy protein, glycinin.

Depending upon pH, ionic strength and other conditions associated with processing, phytate can also form complexes with other components of the soybean including minerals. These conditions, which include extraction pH, temperature, and fermentation with yeast enzymes, may either positively or negatively alter the binding affinity of phytate for minerals. Any reduction in mineral bioavailability may be overcome by the usual measures generally recommended for improving availability from processed plant sources. Fortification is one approach since phytate would not reduce the availability of minerals from other sources. Another approach might be recommending ingestion of 100 mg of ascorbic acid or small amounts of animal protein concurrently with soy products since each improved absorption of heme and nonheme iron in human subjects consuming soy protein (Morck et al, 1982; Lynch et al, 1985). This approach is currently recommended for improving iron absorption from other plant sources such as whole wheat.

Although mineral absorption may be less efficient from soy protein sources compared with animal protein sources, overall mineral balance has not been found to be adversely compromised. Ingestion of soy protein may result in metabolic effects which could actually improve retention of some minerals such as calcium. With lower daily losses, requirements for these minerals are lower. The positive effects of soy protein ingestion on reducing mineral losses should be taken into account when the impact of soy protein on mineral status is evaluated. For example, soy protein is less hypercalciuric than animal protein and does not inhibit vitamin D activation as does phosphate-rich animal protein (Breslau et al, 1988; Portale et al, 1986). Consequently, despite a lower bioavailability of calcium from soy protein, less calcium is lost in the urine and thus mineral balance may not be adversely impacted. In addition, several studies have also indicated that bone resorption may be inhibited by ingestion of soy isoflavones from soy protein sources such as ISP (Potter et al, 1998; Brandi, 1992).

Exposure to Trypsin Inhibitors. Soybeans are a dietary source of trypsin inhibitors. These compounds are protease inhibitors which have been the subject of some debate in the past. The controversial nature of trypsin inhibitors and other protease inhibitors has stemmed primarily from reports suggesting that these compounds may be potent stimulators of pancreatic hyperplasia and hypertrophy in animals, and thus could act as cancer promoters in the presence of chemicals which are carcinogenic to pancreatic cells (Roebuck, 1986). Different sources of soy protein may vary widely in the amounts of trypsin inhibitors they contain, but heat treatment removes most of the activity of these compounds through denaturation (Anderson and Wolf, 1995).

Recent reports have suggested that any previous concern about the association between trypsin inhibitors and risk of pancreatic cancer may have been overstated. Protease inhibitors have been reported to suppress carcinogenesis (Kennedy, 1995), and dietary patterns which include foods containing soy protein have been associated with low rates of a number of cancers (Adlercreutz and Mazur, 1997). Generalizeabiity of data obtained from studies in animals to the human condition may be particularly inappropriate because the animal studies used full fat soy flour and thus the results were likely confounded by the high amounts of fat also consumed from the diets fed to these animals (Erdman, 1981). In addition, there may be species differences in pancreatic of the sensitivity to the proliferative effects of trypsin inhibitors. The size pancreas relative to body weight appears to be a factor determining its proliferative response to trypsin inhibitors. Even among animal models, species with pancreas sizes <0.3% of body weight tend to be less sensitive to the proliferative effects of trypsin inhibitors than those with pancreases comprising a larger percentage of body weight (Leiner, 1979). Since the human pancreas is <0.1% of body weight, it is likely that it would not be sensitive to the effects of the small amounts of trypsin inhibitors found in soy protein (Erdman, 1981). Furthermore, most human pancreatic cancer is ductal in origin (Kennedy, 1995), yet trypsin inhibitors specifically stimulate acinar cellular proliferation (Roebuck, 1986).