Copyright (c) 1995 American Academy of Pediatrics
Pediatrics 1995; 96: 515-517
September, 1995
TITLE: AAP Recommendations on Cow Milk, Soy, and Early Infant Feeding
AUTHOR: FRASER W. SCOTT, PhD, Nutrition Research Division, 2203C, Health Canada,
Sir Frederick Banting Research Centre, Tunney's Pasture, Ottawa, Ontario, K1A
OL2, Canada
TEXT:
The report of the American Academy of Pediatrics (AAP) Work Group on Cow Milk
Protein and insulin-dependent diabetes mellitus (IDDM) states that early
exposure to cow milk protein may be an important factor in the initiation of the
[beta]-cell destructive process in some individuals and recommends that, with
the exception of cow milk-based infant formulas, high-risk infants not be fed
products containing cow proteins during the first year of life. n1 In addition,
the feeding of soy -based formulas was discouraged based on studies reported
from our laboratory. The idea that IDDM might be food-induced has important
possibilities concerning prevention, but the relationship is just not as simple
as first thought. n2
[text deleted]
Interestingly, the part of the report that seems to have attracted the most
attention from your readers was not the linking of cow milk exposure to
diabetes but the cautionary recommendation to avoid soy. Less is known about
soy and IDDM but there are indications that it can be diabetogenic. n10-n13
Although scant, there is evidence that IDDM patients were more likely to have
been fed soy -based formulas than controls n14 and that patients with another
organ-specific autoimmune disorder, thyroiditis, were also more likely to have
been fed soy -based formulas. n15 Diabetes -prone BB rats fed a defined diet
in which the sole source of protein was soybean meal showed a mean diabetes
frequency of 45 +/- 8% standard deviation (SD), considerably higher than the
usual negative control diet with casein or hydrolyzed casein (HC) as the sole
protein source (Figure). Others have also observed that soy -based diets
are
moderately diabetogenic, resulting in 38% n11 and 60% n12 diabetes incidence
in the BB rat.
Soybean meal is a less refined preparation unlike the soy protein isolates
(SPIs) used in soy -based infant formulas. Diets containing SPI were less
diabetogenic in BB rats than soymeal diets; and SPI-based diet produced 33%
diabetes incidence and 25% of animals fed an SPI-based infant formula became
diabetic. However, the rate and final diabetes outcome in soymeal or SPI-fed
BB rats was greater than 2 SD above the mean of negative control, casein, or
hydrolyzed casein diets. Hydrolyzing soy protein did not consistently reduce
diabetogenicity suggesting the diabetogenic activity was retained in smaller
peptides, was affected by varying conditions of hydrolysis, or that the active
component was not protein in nature. Hydrolysis of wheat gluten proteins with
pepsin or trypsin did not reduce the toxicity of this material when fed to
celiac patients but digestion with crude papain was effective. n16 Dr MacLean,
in his letter to the editor, points to the finding by Coleman et al. n6
that a
chloroform:methanol extract of a mainly wheat-based diet was diabetogenic in NOD
mice and suggests that the soy diabetogen is not protein. This is possible but
does not exclude protein as a suspect because wheat, which is diabetogenic in
animals, also contains chloroform:methanol extractable (or modifiable) proteins.
Thus, there is evidence, mainly from studies in the BB rat, that soy diets,
particularly the hexane-extracted soybean meal, which has most lipids removed,
can be diabetogenic. The effect of soy diets in NOD mice is not as clear.
Elliott et al report that a soy protein isolate-based infant formula protected
their low-incidence female NOD mice from developing diabetes, 0% on the
Prosobee diet versus 32% on a cereal-based diet. It will be important to
investigate this finding in high incidence NOD mice maintained under defined,
specific pathogen-free conditions.
[text deleted]
Arly Helm [log in to unmask]
|