Being Vegan and Eating Soy: Myths, Truths, and Everything in Between
by Christa Novelli, M.P.H.
Soy foods have received a great deal of attention in the media in recent years. Very little that the public hears about soy is neutral. Depending upon whom you choose to believe, soy is either a wonder food or the next asbestos. Even among professionals in the field of nutrition and other sciences, there is much confusion about the conflicting information drawn from the countless research articles published each year on soybeans and their derivatives. While it is unlikely that I will address all concerns or cover every study ever conducted on the health effects of soy consumption, I hope to give a clearer picture of what the research regarding soy and human health tells us.
First, we need to have a basic understanding of what is included in the term “soy foods.” Soy foods encompass a wide variety of items including edamame (whole soybeans), soy flour, soymilk, tofu, soy protein isolates, texturized vegetable protein (TVP), soy oil, and fermented soy products such as tempeh and miso. Adding to this list are all of those processed items that are made from these soy products including soy “meats,” soy cheeses, soy “ice creams,” among many other items. One may be surprised to find out how many foods contain some form of soy (soy flour and soy lethicin are often found in cereals, cake mixes, granola bars, and a whole host of other items). Obviously, vegetarians and vegans are not the only Westerners who eat soy products as a regular part of their diet.
As is often the case when a food or medicine is presented as the new “wonder drug”, soy has experienced a backlash in recent years. A few organizations, most notably the Weston A. Price Foundation (WAPF) and the innocuous sounding Soy Online Service (SOS) in New Zealand, have seized upon any negative studies involving soy products (usually the primary isoflavone found in soy, genestein) and are seeking to have soy foods and formula removed from the public marketplace on the grounds that they cause a myriad of problems ranging from cancers to brain atrophy to immune and endocrine system disorders. As one researcher very accurately noted in his response to an editorial appearing in the professional journal,Leukemia, which has been oft cited as “proof” of soy’s deleterious effects, medical journals are no longer a safe forum for scientists and doctors to discuss and debate new hypotheses with one another. These publications, and in particular the abstracts published in these journals, are increasing being accessed by the public especially in the age of the Internet.
For those of you unfamiliar with scientific writing, an “abstract” is a brief summary of a study’s research question, methods and conclusion(s). Abstracts for many scientific articles can be accessed by the public, free of charge, online at the National Library of Medicine’s website, PubMed and also on Med Line. While I used to be a proponent of “doing your own research,” by searching for information on these websites, the more I have seen about how these abstracts are being used and interpreted the less inclined I am to advise the lay public to draw conclusions from abstracts that can be found on the internet. It is very questionable science to start with your conclusion in mind and then search for any and all studies to support said conclusion. Unfortunately, that is exactly what many of the soy detractors appear to have done, aided by abstracts found online.
A good analogy– if I started with the conclusion that eating oranges and broccoli was going to cause cancer, I could actually find a number of articles that showed that a component of these foods (vitamin C) has been shown in some trials done on animals and cell cultures to proliferate tumors. I could then publish articles warning of the dangers of oranges and broccoli and cite these studies as “proof.” Of course, I would be going against the much greater body of evidence that shows the exact opposite effect.
Another problem that I have found in looking at the research cited by the detractors of soy is that one can come to a conclusion that is at complete odds with the research you are citing by reading just the abstract of an article. For instance, the WAPF has a link on their website that purports to show the dangers of soy by citing a large number of research studies with direct quotes drawn from the abstracts of these published articles. One example includes a study of soy formula published in the American Journal of Clinical Nutrition in 1998 . On the WAPF site, pieces of the abstract are quoted stating that the researchers found significantly elevated plasma isoflavone levels (plant estrogens found in the blood stream) of infants fed soy-based formula and that this may exert long-term health effects for these infants. What one would find, if he bothered to read the entire article, was that the researchers found health benefits (i.e. – “effects”) for children with increased levels of isoflavones in their blood stream – the exact opposite of what WAPF is attempting to assert.
Similarly, this same site, as well as a number of others such as theomnivore.com and bullz-eye.com, mentions numerous articles that found that soy consumption by males reduces serum (blood) testosterone levels   . These studies are cited in online articles with titles like: Soy lives up to its reputation as the breakfast of weenies. WAPF quotes one of the abstracts  as testifying that “replacement of meat protein with soyabean protein, as tofu, may have a minor effect on biologically-active sex hormones which could influence prostate cancer risk.” [Emphasis mine] What the term “influence” in the study findings actually meant was reduce, not increase risk. In addressing a similar study , WAPF casts doubt on the authors of the study when it questions why they focused on their findings that consuming soy reduced the risk of prostate disease and arteriosclerosis, but failed to note that “testosterone levels fell in the volunteers eating soy but researchers did not stress this alarming finding in their conclusion.” There was no reason for the authors to consider a reduction in testosterone an “alarming finding.” In fact, the reduction in testosterone levels was likely the reason for the reduction in disease risk much the same as reduced estrogen in the blood stream of women reduces breast cancer risk.
So, what does the research show? I will attempt to answer that question by addressing the various health problems that soy has been alleged to prevent and, conversely, cause or exacerbate. Some of the major diseases and health problems that soy has been associated with (in a positive or negative manner) include breast cancer, other cancers, neurodegenerative diseases and dementia (“brain atrophy”), thyroid disorders, infertility, and disorders or problems specific to males (often referred to as “demasculinization”). Then there is the whole separate issue of soy formula, which I will address on its own.
The argument has been put forth that the phytoestrogens (plant estrogens) contained in soy promote the growth of cancers.
While a few studies have found that animals who are implanted with cancer cells and then fed soy protein isolates show increased growth of the cancer cells, the majority of the research shows soy to have an inhibitory effect on cancer growth. A recent meta-analysis of the literature relating to soy and cancers of the breast, colon, and prostate, found that individuals who consumed soy had a reduced risk of developing all three of these types of cancers .
When results are not in favor of their conclusion, the opponents of soy often call into question the research method used. Meta-analyses have been questioned as a less than valid research method. SOS has stated in their Marketplace Newsletter that meta-analyses are, “a method upon which many in the scientific community frown.”
This same newsletter also claims that meta-analyses that show benefits to consuming soy are flawed because, “there is also the added temptation for researchers, particularly those funded by industry, to omit studies that prevent desired conclusions.” Omission of studies that find the opposite of what one has already concluded is certainly a significant flaw and the conclusions drawn by meta-analyses are appreciably dependent upon proper use of the research tool. A meta-analysis involves conducting a systematic review of the research on a particular topic. If one uses a narrative approach rather than a systematic approach, that is picks and chooses which studies to include based upon personal bias, there is plenty of room for inaccurate conclusions. The key words here being systematic (an organized, unbiased synthesis of the data) versus narrative (an account based upon personal selection).
In fact, properly conducted meta-analyses are one of the more powerful statistical tools for assessing the effect of health interventions and are not “frowned upon” by any scientific community with which I am familiar. The meta-analysis, to which I referred, above, was systematically conducted and showed no apparent bias.
With that aside in mind, women may want to exercise some caution if they have a history of estrogen dependent breast-cancer. While there is no definite proof that consumption of phytoestrogens increases the risk of a reoccurrence, and some studies even show a reduction in risk of reoccurrence, the jury is still out in this area. There is some indication that consumption of soy isoflavones may interact with the anti-breast cancer drug, tamoxifen, reducing its efficacy . While this effect is not certain, it would be wise to discuss your consumption of soy products with your doctor if you have a history of breast cancer.
The most logical course to take if one is concerned about the possible negative effects of plant estrogens contained in soy is to avoid consuming highly processed forms of soy, or reduce consumption of those products. It has been suggested by a few scientists that the degree of processing effects how soy phytoestrogens will react with the human body, with the more highly processed soy products (especially soy protein isolates) exerting a more estrogenic effect in the body. Just to put this in perspective, some of the same researchers who have found estrogenic effects on animals when feeding them high concentrations of soy protein isolates, have also found similar estrogenic effects for isolates of other plants – namely cabbage and brussel sprouts . No one is suggesting that we all stop eating cabbage because it obviously offers significantly more benefits than it does risks. We just generally do not eat extracts of freeze-dried cruciferous vegetables. The same should hold true for soy and any other food – the closer to its whole form as possible is probably the best choice.
Soy causes Alzheimer’s disease, dementia, and reduced brain functioning.
The great majority of the research relating to soy consumption and brain health shows soy to have a protective effect on the brain, not a damaging effect on the brain. One article that has frequently been cited as “proof” that soy damages the brain found some possible benefits and some possible harm for male rats fed a diet high in phytoestrogens . The researchers who conducted this study have also conducted a number of other studies wherein the conclusions were that soy phytoestrogens are protective to the brain.
Edwin Lephart, the lead researcher on the above noted study, has this to say in response to his article being cited as proof of soy causing neurodegeneration of the brain: “In general, phytoestrogens, and more specifically isoflavones, appear to be neuroprotective. Meaning these molecules protect the brain cells and brain function. The research that you state via the Price foundation citing our article is but one study. There are many studies and many molecules that isoflavones influence in the brain and when one examines the overall effect of isoflavones the current evaluation, in my opinion, is that phytoestrogens derived from soy are neuroprotective. ”
Soy consumption reduces thyroid function and/or causes thyroid cancer.
Some older studies of infants fed soy-based formula not fortified with iodine showed reduced thyroid functioning in those infants. Further research has shown that, in the absence of an iodine deficiency, soy does not reduce thyroid function . In relation to infants with congenital hypothyroidism, there is not yet complete consensus amongst the scientific community as to whether soy formula complicates this pre-existing condition. If you have an infant with congenital hypothyroidism, again, soliciting the advice of your doctor or health care provider would be a wise decision.
As far as thyroid cancer is concerned, large population-based studies show that soy not only does not increase the risk of thyroid cancer, it exerts a protective effect .
Infertility and reproductive health
Soy causes infertility, hormonal imbalances, hypospadias, and desmasculinization of males.
A few studies of the offspring of rodents who were fed soy isolates during pregnancy and lactation and rodents who were injected with soy isoflavones in infancy have found evidence of negative reproductive outcomes. However, there is a significant difference in the amount of estrogen to which fetal rats and fetal humans are normally exposed. Fetal rats are normally exposed to very small quantities of estrogen in utero whereas human fetuses are normally exposed to large quantities of estrogen in utero – this is without any soy consumption on the mother’s part. It has been suggested that rats, unlike humans, are not meant to be exposed to significant quantities of estrogen in utero and their reproductive organs are, therefore, more biologically sensitive to the effects of isoflavones. 
Whatever the mechanism for the poor outcomes for the rodents, these results have not been replicated in humans. In fact, research on human infants exposed to soy in utero and during infancy has found no statistically significant differences in sexual development or reproductive health other than one study that found slightly longer menstrual periods in women who were fed soy formula as infants . Clinical studies of infants fed soy formula, arguably the infants with the highest exposure to soy, have found no hormonal defects, no increase in infertility and normal sexual maturation.    
The one human study that implicated a vegetarian diet during pregnancy as a cause of hypospadias  (a congenital defect of the male genitalia in which the opening is not at the tip of the penis) should not be entirely dismissed, but is certainly not conclusive either. The researchers followed a large cohort of pregnant women in Britain and looked at many factors that could influence the development of hypospadias, including maternal diet. They then followed up to determine which factors were associated with a greater instance of hypospadias. Among those factors considered, the only ones that increased the risk of having a male child born with this defect were iron supplementation during pregnancy, being vegetarian during the pregnancy, and/or influenza during the first trimester.
Without looking any further, these findings may be alarming. However, we need to consider what aspect of their environment may have increased the instance of hypospadias among the boys born to the vegetarian mothers – soy is certainly not the only possibility and perhaps not even the most logical suspect once the remainder of the findings are considered. The researchers do hypothesize in the abstract of their article that “vegetarians have greater exposure to phytoestrogens than do omnivores, these results support the possibility that phytoestrogens have a deleterious effect on the developing male reproductive system. ”
We must understand that a hypothesis is only a tentative explanation and not a tested theory. Yes, it is possible that phytoestrogens played a role, but it is also possible that the pesticides on the larger quantities of produce eaten by the vegetarian women played a role, another hypothesis that the researchers offer in this paper. (Very few of the women in the study ate organic produce. Of those who did, none had a son born with hypospadias, but the numbers of women who fell into this category – consuming only organic produce – was so small that it was not possible to calculate if the protective effect of organic produce was statistically significant.)
Again, if we bothered to read the entire article, we would find the statement that, “the association of hypospadias with a vegetarian diet was not obviously explicable by the components of a vegetarian diet. There were differences in the proportion of hypospadias cases born to mothers consuming soya milk or other products, but they were not significant, possibly because there were too few mothers who reported consuming such foods. ” [Emphasis mine] Thus, there was an association with being vegetarian during pregnancy and increased risk in having a son born with hypospadias, but there wasnot an increase in risk related to consuming the two soy products that the dietary questionnaire accounted for: soy milk and soy “meats.”
Without a more thorough analysis of the dietary habits of pregnant women, it is impossible to determine if soy was the cause of the increased incidence of hypospadias among the vegetarian women included in this study or if there was some other dietary or other environmental cause. Further research is warranted, but until these findings are replicated with a more careful dietary analysis, we really do not know what caused the increase in hypospadias in the infants in the study.
Then we have the issue of soy formula. The feeding of soy-based formula to infants is highly contentious. All doctors, scientists and health professionals with whom I spoke (myself included) would recommend breast feeding over any kind of formula when at all possible. With that aside, it is hard to determine the true effects of soy formula on human infants because so few studies have been done on human infants. The Arkansas Children’s Nutrition Center based at the University of Arkansas for Medical Sciences (UAMS) has a research team who is in the process of conducting the largest and most complete study to date of infants fed soy formula, cow’s milk formula and human breast milk. The lead researcher for this study, Tom Badger, was kind enough to answer some questions for me via email and to speak with me on the phone.
Dr. Badger stated that the number of children fed soy formula in the U.S. is larger than in any other nation (25% of the four million infants born in the U.S. each year are fed soy-based formula). This sets the U.S. up as the perfect location for testing the long-term effects of this feeding choice. Dr. Badger estimates that 1 million infants are fed soy formula each year in the U.S. and “thus, we have a lot of experience with this food source for infants… These children grow and develop normally and have disease and disorder rates equal to those of infants fed milk-based formula. All of our research thus far suggests that there are health benefits rather than adverse effects of consuming soy formula. Therefore, any adverse effects would need to be very subtle or occur much later in life, say after age 40-45 years. ”
The Arkansas study currently has hundreds of children enrolled with more children being enrolled every week. The oldest children in the study are three years old at this point and have been followed since four weeks of age. The children in the study are seen every month for the first year of life with each visit lasting between two to six hours. The study participants are seen yearly thereafter until age six, with the hopes that funding will be extended to allow the children to be followed through puberty. Thus far, all evidence from this study (which includes body composition, brain growth and function, endocrinology, and metabolism) shows no difference in the children fed soy-based formula versus cow’s milk-based formula.
Dr. Badger’s team did an in-depth review of prior studies that showed adverse effects on young animals who were injected with or fed soy derivatives. They spoke with the researchers of those studies and concluded that there were major flaws in the studies and/or there were other factors that made the studies inapplicable to human infants. Dr. Badger did note that animal studies are not always flawed by virtue of the fact that they were done on non-human animals and the flaws found varied depending on the studies.
In general, the concern with the animals studies that relate to isoflavones fall into a few areas. One concern is that many of the studies that found increases in cancer dealt with immunocompromised animals and/or female animals whose ovaries and uteruses had been removed. An average adult or child has not had a complete ovo-hysterectomy nor is s/he immunocompromised. It is not possible to generalize to a whole population based upon a small component of that population with different circumstances.
Also, some animal models inject purified isoflavones into rats. Isoflavones found in soy are bound into a completely different form in the whole food (soy milk and tofu are included in this discussion as “whole foods” as the processing involved – heat from cooking and blending of the whole soybeans with water and coagulants – does not turn them into purified isoflavones). Soy consumed as a whole food, including soy milk, tofu and fermented soy products, must be metabolized in order to extract the isoflavones. Feeding or injecting an animal with purified isoflavones skips over this metabolic process and is basically the same as giving the animal high doses of a drug.
Unfortunately, animal studies that include feeding soy products (not just purified isoflavones) cannot always be easily translated into human effects either. For instance, studies that have found reduced testosterone levels in male monkeys fed soy formula, failed to take into account the different intestinal bacteria found in monkeys that are not found in humans. Despite our close genetic relation, monkeys have bacteria in their intestines that metabolizes daidzein (a soy isoflavone) into a highly estrogenic substance; humans do not have this same bacteria.
The animal models can be a good place to start, but should not be relied upon for making health decisions for humans without replication in human subjects. At this point, those studies that have been conducted on humans, and even the majority of those conducted on animals, have leaned towards the conclusion that soy is beneficial to health, not detrimental.
What forms of soy and how much soy should you be eating?
Since most of the studies that raise concerns about soy consumption seem to focus on pre-metabolized isolates or other similarly processed forms of soy, the wisest choice to is to choose to eat soy foods as part of a balanced diet and in the least processed form available. Soy can be a valuable source of protein in a vegetarian or vegan diet, but it should not be relied upon as the sole source of protein. We should also be careful with drawing the conclusion that, if a little bit of something is good, a lot must be even better. Variety in your diet is an excellent way to ensure balance and good health.
Choosing traditional and less processed forms of soy (such as tofu, miso, tempeh, edamame and even soy milk) over highly processed soy foods (such as soy cheeses, soy meats, and soy protein isolates) is likely to be a safer choice, as well. We all need our treats sometimes and the research does not support the conclusion that eating a little vegan soy cheese on your pizza is going to cause health problems, so go ahead and allow yourself some indulgences every now and then if you want to. Just don’t make soy cheese and soy sausage casserole the mainstay of your diet.