Soy: To Fear or Not to Fear?
Learn what the latest science says about soy safety — and why we were misled.
A dear friend of mine, who happens to be a cancer survivor, loves soy lattes, yet is afraid to indulge. Why?
Like many cancer survivors, she worries that the phytoestrogens will fuel tumour growth.
It’s not just cancer survivors that fear soy. To get a sense of the scope of soyaphobia, simply head to the health conscious shopper’s mecca: Whole Foods. You’ll quickly see that the soy-free label rivals the GMO-free label in its popularity. Notably, soyaphobia doesn’t seem to be nearly as common in Germany, where I’m sipping a soy latte as I finish this article (it’s strange not to see nut milk — a favourite North American soy alternative on any café menus!). Some day I’ll look up some real data on this epidemic…
What do soy-avoiders fear? Women worry it will cause breast cancer or reproductive issues, while men worry about ‘man boobs’ or lower sperm count. Others don’t know exactly what they fear but would rather “play it safe”.
Yet, the latest science, based on dozens of studies in thousands of people, tells us loud and clear that enjoying soy is safe, and may even have protective health benefits.
This article tells the story of soyaphobia, one that provides insights into the pitfalls of science and its communication to the public. By the end of the story, I hope you will join me in fearlessly embracing this fabulous food.
Why should you embrace soy?
Soy is one of the best plant-based sources of protein, fiber, and more. It’s also delicious, versatile, easy on the planet and budget-friendly. Read more here.
Where did soyaphobia begin?
The story of soyaphobia began in the 1940s when an Australian sheep farmer noticed fertility issues in his flock. He identified clover, one of their favourite foods, as the culprit. It turns out that the clover was rich in phytoestrogens.
Phytoestrogens are plant-derived chemicals that have a similar shape to the human hormone estrogen. There are several classes of phytoestrogens, including isoflavones (found in clover and soy).
A theory was born. If eating phytoestrogens causes fertility issues in sheep, then eating phytoestrogens may cause similar issues in humans. The spotlight landed on soy because it is one of the richest phytoestrogen sources in the human diet.
If soy phytoestrogens indeed acted like a dietary estrogen boost, we would expect to see:
- fertility issues similar to those seen in sheep
- higher rates of breast cancer
- higher rates of disease recurrence in breast cancer survivors
- negative impact on male sex hormones
On the plus side, we would expect soy to help with conditions related to low estrogen, such as a hot-flash treatment for women going through menopause or a bone-density loss in post-menopausal women.
Were these predictions true? It took several decades to get the data we needed — large human trials — to answer these questions confidently.
Spoiler alert: Soy is safe!
Large human trials of soy safety
Here are just a few of the large studies that allow us to declare soy as safe, and to do so with confidence:
- Breast cancer (2009): Women who consumed moderate soy had the same or lower rates of cancer recurrence than those who avoided soy, even among those who had cancers that were estrogen-driven. See this study of over 5000 breast cancer survivors and this study of nearly 2000 survivors.*
- Prostate cancer (2017): A meta-analysis of 22 studies (over 10,000 cases and 100,000 controls) study found no greater risk of prostate cancer associated with soy phytoestrogens (and some hints of benefits). (Study)*
- Men’s hormone health (2010): A meta-analysis of 15 clinical studies found no effects of soy protein or isoflavones on reproductive hormones in men (Study).
- Soy infant formula (2001): Soy formula has a long track record of safety, including no impact on reproductive health into adulthood (Study).
I am not aware of any large, high quality human trials that have shown a harmful impact of soy consumption (please tell me if I missed any!!). I’ll elaborate shortly on the some of the studies that stoked the fires of soyaphobia. Spoiler alert: They have severe limitations that make their our confidence in their conclusions weak at best.
Note: The largest body of evidence for soy safety uses levels up to 100 mg of isoflavones (phytoestrogens) per day which corresponds to about 3–4 servings of soy or 25–30 grams of soy protein. Some supplement studies use twice this level.
Additional evidence: Asia as a natural experiment.
In a perfect world, we could turn to “gold standard” studies to answer our health questions: large, placebo-controlled randomized trials. Unfortunately, such studies are rarely done outside of the pharmaceutical testing because they are very expensive, risky, and time consuming — and in some cases, unethical (especially for a suspected toxins). When placebo-controlled clinical trials are lacking, we turn to observational studies, including “natural experiments” (and some of the case-control experiments above).
In a “natural experiment” we seek signs of benefit or harm in those that naturally consume a certain food. This approach is powerful for “exonerating” a food that faces an accusation of toxicity or harm. If, for example, tea-drinkers were found to have lower stomach cancer rates, it’s would be hard to argue that drinking tea causes stomach cancer.
Asian populations are the world’s highest soy consumers, yet they show no evidence of the predicted harms. In fact, some disease rates are the opposite of what you would predict if phytoestrogens boosted your own estrogen. For example, rates of breast cancer are historically lower in Asia than in North America (see article). Typical daily Asian consumption is 6–11 g of soy protein (25–50 mg of isoflavones), but some Asians approach 25 grams per day (100 mg isoflavones) (see article).
It’s important to note that observational studies are much less reliable for proving guilt than innocence. If tea drinkers had higher rates of stomach cancer, you CANNOT infer that drinking tea causes stomach cancer. As you have probably heard, correlation does not equal causation because two things are often related for some hidden reason. The lack of symmetry between testing innocence versus guilt stems from the fact that correlations are (generally) necessary — but not sufficient — to prove causation. Read more here about evaluating causal claims from a logical perspective.
Why were we misled?
At this point you may be wondering how fear of soy could have become so rampant given the vast evidence for safety. It can’t be all due to infertile sheep, can it? The answer lies in over-zealous interpretation of early research on soy phytoestrogens, together with our natural tendency to adopt and hold on to fears.
Human health research follows a predictable path: 1) preclinical “model systems” (such as rats and lab-grown cells); 2) small human trials; 3) large human trials. Each phase produces more reliable findings, yet is dramatically more expensive, time consuming, and risky than the last. Findings from earlier steps are used to inform and “gate” the next phase of experiments.
In the case of soy, a handful of early studies raised red flags and fueled the fires of soyaphobia despite their clear limitations. For example, this study fed rats massively unrealistic amounts of soy, and this study fed purified soy proteins to lab-grown breast cancer cells. Small, uncontrolled human studies were also done — such as this study with no placebo group (invalidating its fearful conclusions).
The soy saga shows what happens when we tout findings in model systems and small human trials as conclusive, when in reality they are highly fallible — a fact that even the strongest advocates of “model systems” acknowledge. Unfortunately, the press tend to ignore this nuance and encourage us to latch onto fears rooted in “very preliminary” evidence.
The bottom line on soy safety
Based on the latest scientific evidence, my family and I happily consume 2–3 servings of soy daily. We are good company in reaching this conclusion (see Appendix for expert opinions).
Three servings of soy foods per day (on average), with a focus on natural sources (soy, such as tofu, tempeh, soy milk, and edamame) aligns with the highest Asian consumers, and the largest body of scientific evidence. That said, don’t fret about the occasional soy-a-rama.
What about kids? While there isn’t as much data, I find the safety record of soy formula to be extremely comforting, as this delivers a higher dose per body weight than adults normally get. This leaves me comfortable with between 50–75mg per day of isoflavones (2–3 servings). We are in the clear with 2 cups of soy milk plus a serving of tofu or edamame per day.
I contacted the manufacturer of our favourite soy milk (Silk) to ask about isoflavone levels. They told me:
Each gram of soy protein naturally contains about 2–3mg of isoflavones; about 12–24 mg of isoflavones per each 8 oz serving depending on the flavor. This goes for both our U.S. and Canadian products!
Since science is never a closed book, I should say that we can never be sure that there isn’t a subset of people that suffer some harm that we have yet to discover — but we could say the same for any food.
Enjoy that soy latte, Dana!
Read more about soy’s winning combination of delicious, nutritious, versatile, environmentally friendly, and wallet-friendly!
Closing thoughts on food fears
It’s natural to play it safe when a food fear hits the news. However, I urge you to take any health headlines based on model systems with a grain of salt. Similarly, small, uncontrolled human trials should be interpreted very cautiously — especially when they make claims of guilt.
Just as important as being leery of adopting new food fears is learning to ditch fears that no longer make sense. To do this, we must be open to new, stronger evidence, regardless of whether or not it supports the position you hold. This is easier said than done. I highly encourage you to read “Mistakes Were Made But Not By Me” for a psychological look at our tendancy to cling tightly to our beliefs even in the face of evidence to the contrary (by social psychologists Carol Tavris and Elliot Aronson).
Stay tuned for more on letting go of food fears, becoming a smarter consumer of health headlines, and understanding the pitfalls of science.
Please also let me know if you would like to learn more about the science behind claims that soy offers additional health benefits beyond its fabulous nutritional profile.
I am formally trained in human genetics (PhD) and spent the first decade of my career working in cancer research, drug development, and personalized medicine.
This new chapter of my career is dedicated to helping others to live their healthiest lives, using an evidence-based paradigm. I strongly advocate for a plant-rich diet and love helping others to become #plantsmart — and in this case #soysmart.
For more nutritional insights — and healthy, simple, recipes (including many with soy) — check out my website (Fueled by Science).
Expert positions on soy and health
- American Institute of Cancer Research
- MD Anderson
- Linus Pauling Institute
- Mayo Clinic
- Cleveland Clinic
- Physicians Committee for Responsible Medicine
… and more