The Story of Soy Formula Feeding in Infants: A Road Paved With Good Intentions

To the Editor: Three documents published over the past few months have once again drawn the attention of the scientific community to the controversy regarding the use of soy-based formulas in infant feeding and have placed their associated pitfalls squarely in the forefront of public and specialist medical concern. These publications are the American Academy of Pediatrics' recommendations for the use of soy protein-based formulas in infant feeding (1), a report on the effect of phyto-oestrogens from soy-based infant formulas by Setchell et al. (2), and the report presented to the European Parliament by Jensen on July 23, 1998 (3). The American Academy of Pediatrics' recommendations state that 1) isolated soy protein-based formulas have no advantage over cow's milk protein-based formulas as a supplement for the breast-fed infant, 2) they are appropriate for use in infants with galactosaemia, hereditary lactose deficiency, and children of vegetarian parents, 3) they have no proven value in the prevention or management of infantile colic, nor in the prevention of atopic disease in healthy or high-risk infants, and 4) infants with cow's milk protein-induced enteropathy or enterocolitis should not be given isolated soy protein-based formulas routinely, whereas they can be used in documented IgE-mediated allergy to cow's milk protein. Setchell et al. (2) provide convincing evidence that infants fed soy-based formulas are exposed to isoflavones to a 6- to 11-fold higher extent on a body weight basis than occurs at the dose that has hormonal effects in adults consuming soy foods, and this may be sufficient to exert biological effects. As if this evidence were not convincing enough, Jensen (3) very recently presented a proposal for the regulatory control of all phyto-oestrogens present in foods because they are potentially harmful, particularly for women and infants. All paediatricians should therefore exercise the utmost caution in prescribing soy-based infant formulas. As pointed out by the American Academy of Pediatrics, such formulas were first introduced in infant nutrition in 1909 by Ruhräh (4), who 2 years later (1911) went on to advocate their use in summer diarrhoea (5). In 1929, Hill and Stuart (6) proposed soy protein-based formulas for the treatment of cow's milk intolerance. The idea was sound because soy flour is less expensive than cow's milk and has a sufficient nutritional value. Unfortunately, in the 1970s the infant food industry realised that selling soy-based formulas at a high price as "special formulas" for various infantile diseases such as acute gastroenteritis and cow's milk intolerance could be extremely lucrative. As a result of this trend, in several European countries and particularly in Italy, the use of soy-based formulas became common and much more widespread than was strictly necessary. Indeed, when one of the authors (SG) was asked in 1990 to review the evidence obtained from 33 Italian pediatric gastrointestinal centers dealing with infants and children with a proven diagnosis of cow's milk protein intolerance, it was found that soy-based formulas were chosen as the substitute for milk-containing formulas in about 50% of 2108 children. Despite this, soy-based formulas were clearly those associated with the highest rates of failure when compared with both protein hydrolysates and "home-made" diets. They had to be discontinued in 47% of cases overall, ranging from 53% of infants younger than 3 months of age to 35% of children older than 1 year of age. The first serious misgivings were voiced in 1983 by the American Academy of Pediatrics (7), whose members pointed out that such formulas should not be used in premature or low-birth-weight infants, in the dietary management of documented clinical allergic reactions to cow's milk protein or soy protein formulas, or in the routine management of colic. In the same year, Zoppi et al. (8) published an article in which they demonstrated that the best response to vaccinations is obtained in healthy breast-fed infants and the worst in healthy infants fed soy protein-based formulas. This article followed two previous reports in 1982 (9) and 1979 (10) in which Zoppi et al. found evidence of abnormal gammaglobulin, immunoglobulin, transferrin, complement fractions, and T-lymphocyte markers in healthy infants fed soy protein-based formulas. They later reported similar results in 1986 with an experimental oral rotavirus vaccine (11). The article published in 1982 (9) was described as innovative in an editorial in the Journal of Pediatric Gastroenterology and Nutrition(12). None of these papers, however, received the attention they deserved from another Italian group (13) who studied not healthy infants but babies of atopic families in which the antibody response to vaccinations may be impaired. It was a classic case of "Nero fiddling while Rome burned." While we were busy haggling over these important issues, the consumption of soy-based formulas in Italy increased from 160.7 tons in 1982 (= 3.4% of cow's milk formulas) to an astonishing 688.9 tons in 1990 (= 9.1% of cow's milk formulas), only to drop back to the still very substantial figure of 350.4 tons in 1997 (= 5.7% of cow's milk formulas), with a total annual outlay for Italian families of around 20 billion lire ($12 million [U.S.]) (14) (Fig. 1). During the same period the consumption in the United States has been around 25% of that of cow's milk-based formulas (15). This is clearly big business for the industry, which can hardly adduce effective medical necessity in partial justification of its very high returns. Finally, if we consider the potential allergenicity of soy-protein formulas first described in 1934 by Duke (16) and later confirmed by several other authors (Table 1), we should stress that the issue remains a bone of contention between those who are in favour of soy-protein formulas (13) and those who are against them (35). This divisive clash of opinion is also reflected in the mutually antagonistic stances adopted by two important scientific societies, the European Society for Pediatric Gastroenterology, Hepatology and Nutrition (37) and the European Society of Pediatric Allergy and Clinical Immunology (38). The issue is also enmeshed with the question of the use of hydrolysed cow's milk formulas in the prevention of allergies to cow's milk, but the subject has not been addressed in a scientifically dispassionate manner (39).FIG. 1: Percentage of soy-based formulas sold in Italy as a proportion of total cow's milk-based formulas and overall costs for the families (14).TABLE 1: Authors who demonstrated the high incidence of allergenicity of soy-protein formulasBe this as it may, in the light of the latest documents, we can conclude that the story of soy-based infant formulas started out with the best of intentions but has now turned distinctly sour. These formulas should no longer be extensively used, at least not in developed countries, where protein of animal origin is not particularly expensive and is widely available. Acknowledgment: The authors thank Anthony Steele, MA, Lecturer in Medical English, University of Verona, for his invaluable linguistic editing of this letter. Giuseppe Zoppi Stefano Guandalini Division of Pediatrics, University of Verona, Verona, Italy. Department of Pediatrics, Section of Gastroenterology, Hepatology and Nutrition, Chicago, Illinois, U.S.A.