You have probably avoided soy protein because someone told you it messes with your hormones. A dose-response meta-analysis tested that fear at every dose from 49 to 164 mg of isoflavones per day in 329 men. The measure of disagreement between studies: zero percent.
The measure of disagreement between the 8 studies that measured testosterone: zero percent.
You have probably turned down soy protein at least once. Maybe a friend told you it lowers testosterone. Maybe you read it in a listicle titled "Foods That Kill Your T Levels." Maybe you just absorbed the idea the way most gym-goers do, from everywhere and nowhere, a quiet certainty that soy is something men should avoid.
A research team wanted to know what actually happens. They pooled every qualifying randomized controlled trial ever published on soy and male hormones. Eleven studies, 489 men, five countries, tested between 2000 and 2018.
Then they did something no previous meta-analysis had attempted: they built the dose-response curve. Not just "does soy affect testosterone" but "at what dose does anything happen?"
The measure of disagreement between the 8 studies that measured testosterone: zero percent. Every research team got the same answer.
Soy does not lower testosterone at any dose tested in human trials, and the one hormone that does respond has a threshold above what food provides.
- Testosterone showed no response to soy regardless of how the researchers divided the data, by age, study design, exercise status, supplement form, or dose level. The null finding is comprehensive, not just average.
- Estradiol, the primary form of estrogen in men, did respond to dose, but the threshold sits above what food-based soy provides and below what most supplements deliver.
- The belief that soy lowers testosterone traces to a 2003 rat study using a species that processes soy compounds differently from humans, amplified by internet culture from 2017 onward.
- All studies lasted 4 to 12 weeks. Whether the testosterone finding holds over years of daily soy consumption remains an open question with only indirect reassurance from population data.
The Most Unanimous Finding on Soy and Testosterone
Eight randomized controlled trials measured testosterone in 329 healthy men. The men consumed soy products or soy isoflavone supplements. The doses ranged from 49 to 164 milligrams of isoflavones per day. The interventions lasted 4 to 12 weeks. The study sites spanned the United States, Japan, Australia, the United Kingdom, and Canada.
When all eight studies were pooled, the average difference across 329 men was so small it could have pointed in either direction. Soy moved testosterone by an amount indistinguishable from noise.
But the number that matters most here is not the effect size. It is the I-squared statistic: 0.0%. I-squared measures how much studies disagree with each other. A value of zero means every study pointed the same direction by the same amount.
Not "most studies agreed." Not "the trend was consistent." Every team that ran this experiment, in every country, across every study design, using different soy products at different doses for different durations, got the same result.
Then they checked the subgroups. Parallel studies versus crossover studies: no effect. Men under 40 versus men over 40: no effect. Studies with exercise versus without: no effect.
Soy supplements versus soy foods: no effect. Higher isoflavone doses versus lower: no effect. Every way they sliced the data, testosterone did not respond to soy.
What Happens When You Triple the Soy Dose
Prior meta-analyses on this topic reached the same verdict. Reed 2021 pooled 41 studies. Hamilton-Reeves 2010 pooled 15. Both concluded that soy does not affect testosterone.
But they answered a simpler question. They asked whether soy has an effect on average. Rajaie's team asked whether soy has an effect at any dose.
They built a dose-response model. It covered 306 participants across 7 studies. They used three different mathematical approaches. They tested whether the relationship between soy isoflavone dose and testosterone was linear, curved, or wavy.
Every model came back flat. The straight-line model found nothing. The curved model found nothing, and the wavy model found nothing. Three mathematical approaches, and not one found a relationship between soy dose and testosterone.
The man in the supplement aisle wondering "maybe a little soy is fine, but what if I take too much?" now has an answer.
From 49 milligrams of isoflavones per day to 164, roughly the difference between a glass of soy milk and multiple servings of concentrated soy protein, testosterone did not change at any point along the curve. There is no "too much" for testosterone within the entire range these studies tested.
Every way they sliced the data, testosterone did not respond to soy.
How a Rat Study Became a Gym-Floor Fact
The fear did not come from nowhere. In 2003, researchers fed high-dose genistein (a soy isoflavone) to infant male rats and observed reproductive changes. That finding was real, in rats.
But rodents process soy compounds through a fundamentally different pathway than humans. The result was never replicated in human clinical trials.
The path from there is traceable. The term "soy boy" emerged on 4chan in 2017, built on the idea that soy consumption reduces masculinity. It spread to mainstream internet culture, TikTok, Reddit fitness forums, and eventually the vocabulary of gym floors everywhere. The insult's entire scientific basis is a species that processes soy differently than we do.
You did not choose this belief. It was handed to you by a culture that inherited a finding from the wrong animal model and never checked it against the pooled human data. This meta-analysis did the checking: 11 randomized controlled trials, 489 men, and a flat line where the fear predicted a cliff.
The One Male Hormone That Actually Responds to Soy
Most coverage of soy and hormones stops at testosterone. This study measured four hormones, and one of them tells a more complicated story.
Estradiol, the primary form of estrogen in men, did respond to soy isoflavone dose, but not in a straight line. As the dose increased, estradiol levels rose, peaking at approximately 6.5 pg/mL above baseline when intake reached about 72 milligrams of isoflavones per day. Above that dose, the curve turned back down.
Before the anxiety lands, three layers of context.
First, where that threshold sits. Traditional soy foods provide about 3.5 milligrams of isoflavones per gram of soy protein. To reach 72 milligrams per day from food, you would need roughly 20.6 grams of soy protein from traditional sources daily. That is a soy-heavy diet.
Japanese older adults, the population with the highest lifelong soy intake, typically get 30 to 50 milligrams of isoflavones per day. Comfortably below the threshold.
Second, the size of the shift. Even at the peak dose-response point, estradiol moved by about 6.5 pg/mL. The normal male range for estradiol is 10 to 40 pg/mL. A shift of 6.5 on a 30-point scale, at a dose most people never reach from food, is not the hormonal crisis the internet warns about.
Third, the shift might be smaller than measured. Most control groups in these studies used whey or casein rather than a true placebo. The study's own authors flag this: the estradiol increase "may be biased by the latent diminishing effect of high-quality protein sources used as comparator groups."
If the control protein independently pushed estradiol down, then part of the measured "soy effect" is actually a control-protein effect. The real soy contribution to estradiol could be smaller than 6.5 pg/mL.
The case reports that drive the most visceral version of the fear involved intake approximately 5 to 9 times typical Japanese levels. One widely cited case involved a man consuming roughly 3 liters of soy milk daily.
These cases are consistent with the dose-response curve, not contradictory. They confirm the threshold exists. Below it, nothing measurable; way above it, problems. The threshold sits above what food provides.
The other hormones told smaller stories.
SHBG, the protein that binds testosterone in the bloodstream, followed a similar curve to estradiol, peaking at roughly 9 nmol/L above baseline around 120 milligrams of isoflavones per day — but only 3 studies with 119 participants measured it, and a single data point supported the high end of the curve. Free testosterone showed no dose-response relationship at all.
The contrast between these two findings is what makes this page honest. Testosterone: bulletproof. Estradiol: real threshold, above typical food intake, fragile evidence base, potentially confounded measurement.
Where the Evidence Gets Honest
The testosterone finding is as solid as nutrition research gets. Eight studies, zero heterogeneity, consistent across every subgroup. Confirmed by two prior meta-analyses and backed by a flat dose-response curve. An independent evidence-quality assessment rated the certainty as moderate.
The estradiol finding is fragile.
Four studies measured it. Remove one of them, Haun 2018, and the overall result flips from no significant effect to a significant positive effect. One study changes the conclusion. With a pool this thin, the overall result cannot be considered definitive.
The evidence-quality assessment actually rated the estradiol dose-response pattern as high certainty — not because the exact size of the shift is settled, but because the pattern itself is unmistakable. Estradiol does respond to dose. How much it responds above 70 milligrams per day is where the uncertainty lives.
Every study lasted 4 to 12 weeks. Nobody has run a years-long controlled trial on soy and testosterone. The researchers say so directly — they want longer studies. Populations in East Asia have consumed soy for generations without population-level hormone problems, which is reassuring but not the same as a controlled experiment.
The contrast between these two findings is what makes this page honest. Testosterone: bulletproof. Estradiol: real threshold, above typical food intake, fragile evidence base, potentially confounded measurement. Both statements are true. Both leave you better informed than the person who told you to avoid soy.
The Numbers That Replace the Fear
The belief that soy lowers testosterone arrived without numbers. It was vague, culturally inherited, and never examined. This meta-analysis replaces that vagueness with precision.
The next time someone says soy lowers testosterone, the response is not "I read somewhere that it doesn't." It is eight studies, five countries, zero percent disagreement, flat at every dose. Precision replaces fear.
And now the question that naturally follows: if soy protein does not affect your hormones, can it actually build the same muscle as whey? Forty-three randomized trials tested that too.
You are standing in the supplement aisle deciding between soy and whey protein. The testosterone question is settled: it does not move at any dose these studies tested, and that result holds whether you exercise or not. If your soy intake comes from food (tofu, edamame, soy milk), your isoflavone dose sits well below the threshold where estradiol even begins to respond. The fear that kept soy off your shelf was borrowed from a study done on a different species.
What this means for you
Traditional soy foods deliver roughly 3.5 milligrams of isoflavones per gram of soy protein. A serving of tofu or a glass of soy milk puts your daily intake well within the range where this meta-analysis found no hormonal effect of any kind: not on testosterone, not on estradiol.
Even Japanese populations with lifelong daily soy consumption typically take in 30 to 50 milligrams of isoflavones per day, which sits below the 72-milligram threshold where estradiol begins responding.
Soy protein supplements vary in isoflavone content depending on how they are processed. Soy protein isolate, the most common form in protein powders, typically contains fewer isoflavones than soy protein concentrate. Testosterone showed no response across the entire dose range tested in these trials, including supplement-level doses up to 164 milligrams of isoflavones per day.
The estradiol threshold sits around 72 milligrams of isoflavones per day. Whether your supplement reaches that depends on the product, not on the protein dose itself.
The exercise subgroup analysis adds a layer specific to your situation. In studies where participants trained, soy showed no significant effect on estradiol: the one hormone that responded to dose in the overall analysis. In studies without exercise, estradiol did increase.
The between-group difference was not large enough to be conclusive, and only two studies included exercise, so this is a directional signal rather than a settled answer. But for the gym-going soy consumer, it points in a reassuring direction.
Before you change anything
This meta-analysis studied healthy adult males aged 18 and above, consuming soy at isoflavone doses from 49 to 164 milligrams per day, for 4 to 12 weeks. The trials included men from five countries: the United States, Japan, Australia, the United Kingdom, and Canada. These findings do not apply to women, children, adolescents, professional athletes, or men with existing hormonal conditions. All were excluded from the study design.
Every study lasted 4 to 12 weeks. No trial tested what happens beyond three months of daily soy consumption. The estradiol analysis drew from only four studies with 138 participants. The binding-protein analysis drew from three studies with 119 participants. The dose-response curves at higher isoflavone doses are supported by just one or two data points each, with wide margins of error. The researchers specifically call for longer, better-powered trials.
Testosterone evidence is near-bulletproof: eight studies, zero disagreement between them, a flat dose-response curve, and consistent findings across two prior meta-analyses that pooled 15 and 41 studies respectively. Estradiol evidence is fragile. The dose-response threshold is rated high-certainty by the quality assessment (because the gradient was detected), but the measurement rests on four studies, is sensitive to a single study’s removal, and may be inflated by the choice of control protein.
The testosterone question is resolved. But the gym-goer who just cleared that fear has an immediate next question: does soy protein actually build the same muscle as whey? A separate meta-analysis collected forty-three randomized controlled trials, including seventeen that directly compared soy protein to whey protein in the same participants. The verdict on muscle outcomes might surprise you less than the hormone data did, but the scale of the evidence is worth seeing.
What This Study Found
All findings from this paper, in plain language.
- Soy did not change testosterone in any of eight trials testing 329 healthy men across five countries over 4 to 12 weeks.
- Free testosterone stayed the same in three trials with 119 men consuming soy, though the evidence base is thin.
- The protein that carries testosterone in blood was unaffected overall, but the three studies that tested it disagreed substantially with each other.
- Estradiol showed no overall change when four studies were pooled, but this average hides important dose-related patterns.
- Testosterone stayed flat at every dose tested, from moderate to high soy isoflavone intake, with no linear or curved relationship detected.
- Estradiol rose modestly at higher soy doses, peaking around 72 milligrams of isoflavones per day, above typical food-based soy intake.
- A testosterone-binding protein nudged up at very high soy doses around 120 milligrams per day, but only three studies measured this.
- Free testosterone showed no dose-related pattern across the full range of soy isoflavone intake tested in three studies.
- Estradiol increased in sedentary studies but not in exercise studies, though only two exercise studies existed and the difference was not statistically reliable.
- The estradiol result depends on one study. Removing it flips the overall finding from no effect to a significant increase.
- Evidence certainty is high for the estradiol dose-response and moderate for testosterone, free testosterone, and the binding protein.
- Control groups used whey or casein instead of a true placebo, which may have inflated the apparent soy effect on estradiol.
- Traditional Japanese soy intake sits below the estradiol threshold at 30 to 50 milligrams of isoflavones per day, short of the 72-milligram inflection point.
- The testosterone-to-estradiol ratio could not be analyzed because only two of eleven studies reported it.
- No sign of publication bias was found for any of the four hormones tested.