Does soy affect your hormones?

Soy isoflavones look like estrogen on paper. They don’t behave like it in the body. Three decades of clinical data — including a meta-analysis of 40 randomized controlled trials enrolling 3,285 participants — find no meaningful effect of typical soy intake on circulating estradiol, testosterone, or thyroid hormones in healthy adults (Viscardi et al., 2025).

The confusion is understandable. Genistein and daidzein, soy’s primary isoflavones, share structural similarities with 17beta-estradiol. But structural similarity is not pharmacological equivalence. Isoflavones are selective estrogen receptor modulators (SERMs), not estrogens — a distinction with real clinical consequences.

The short answer

• Estrogen in women: no significant change in circulating estradiol, FSH, endometrial thickness, or vaginal maturation index across 40 RCTs (Viscardi et al., 2025).
• Testosterone in men: no change in total testosterone, free testosterone, SHBG, LH, or sperm quality at any dose studied in humans (Hamilton-Reeves et al., 2010; Reed et al., 2021).
• Breast cancer risk and survival: no elevated risk in prospective cohort data; overall survival is modestly better in soy consumers than non-consumers (Qiu & Jiang, 2019).
• Thyroid: no clinically significant change in free T3 or T4 in iodine-replete adults; a small TSH rise appears in some trials but is not considered meaningful (Otun et al., 2019).
• One real caveat: concurrent iodine deficiency may amplify any TSH effect. Adequate iodine removes this concern.

Why isoflavones are not estrogens

17beta-estradiol, the primary human estrogen, binds with high affinity to both estrogen receptor subtypes (ERalpha and ERbeta) and drives cell proliferation in breast, uterine, and other estrogen-sensitive tissues.

Genistein binds preferentially to ERbeta and does so with much lower affinity than estradiol — by several orders of magnitude, depending on assay conditions (Carbonel et al., 2020). The pharmacological class this places soy isoflavones in is SERMs, the same class as tamoxifen. Whether a SERM is “estrogenic” or “anti-estrogenic” depends on the receptor subtype and the tissue. In breast tissue, isoflavones may actually compete with endogenous estradiol at ERalpha receptors, potentially reducing net estrogenic stimulation.

This receptor-level nuance explains why population-level data on breast cancer risk show no elevated signal — and may explain the modest protective association seen in some survival studies.

What the data say for men

The fear that soy lowers testosterone or causes feminisation in men comes largely from case reports involving extraordinary intake — multiple litres of soy milk daily over months. These cases are outliers, not guides to normal dietary exposure.

Meta-analyses at realistic doses are unambiguous. Hamilton-Reeves et al. (2010) pooled 15 placebo-controlled treatment groups and found no effect on total testosterone, free testosterone, SHBG, or estrogen in men. Reed et al. (2021), the most recent meta-analysis on this question, confirms every finding. No published RCT documents gynaecomastia or clinical feminisation from soy foods at typical dietary intake.

This also matters for the claim that vegan men are weaker because of soy. Testosterone levels are unchanged; resistance-trained vegans achieve equivalent strength gains to omnivores when total protein intake is matched — a finding covered in the protein for vegan athletes article.

Soy and breast cancer

Oncologists have historically advised breast cancer patients — particularly those with ER-positive tumours — to avoid soy, on the theory that isoflavones might fuel oestrogen-sensitive cancer cells. The clinical evidence does not support this advice.

Prospective cohort data do not show elevated breast cancer risk from soy or isoflavone intake. A systematic review and meta-analysis by Qiu and Jiang (2019) found no increased incidence risk across populations and reported modestly better outcomes in soy consumers: overall survival HR 0.84 (95% CI 0.71–0.98), a statistically significant reduction. Breast cancer-specific survival showed a similar trend (HR 0.89, 95% CI 0.74–1.07). Higher lifetime soy consumption, common in Japan and China, correlates with lower incidence in population studies, though dietary and lifestyle confounders make causal inference difficult.

These are observational findings — RCT evidence on recurrence is lacking. But the data are at minimum consistent with soy being safe after a breast cancer diagnosis. The oncology consensus against soy is not well supported by current evidence, and blanket avoidance advice warrants scrutiny.

The thyroid question

Soy’s effects on thyroid function are the area where a real, if narrow, concern exists. Otun et al. (2019) found modest TSH elevation across trials, but no significant change in free T3 or free T4 — the hormones that actually determine thyroid activity. A small TSH change without hormone level change is not clinical hypothyroidism.

The relevant caveat comes from Messina and Redmond (2006): the TSH effect appears meaningful primarily when iodine intake is concurrently deficient. Iodine deficiency is worth monitoring on any vegan diet — not because of soy specifically, but because dairy is a major iodine source for omnivores and many plant milks do not fortify with iodine. Eating soy with adequate iodine from iodised salt or seaweed eliminates the thyroid concern.

Why animal studies don’t settle this

Much of the original alarm about soy and hormones came from rodent studies. These studies used doses of isoflavones that are supraphysiological relative to any realistic human dietary intake, and they ran in an animal model that metabolises isoflavones differently from humans.

Rodents convert daidzein to equol at rates above 80%. Equol is a more potent phytoestrogen than daidzein itself. Only about 25–30% of Western humans are equol-producers — the rest lack the gut microbiome capacity to make the conversion (Carbonel et al., 2020). Rat study results cannot be straightforwardly applied to human diets, and they should not be.

Practical guidance

• Eat soy foods without concern at typical serving levels: one to three servings per day (tofu, tempeh, edamame, miso, soy milk) is within the range of populations with decades of safe use.
• Ensure adequate iodine — this applies regardless of soy intake. Use iodised salt, or include seaweed occasionally.
• Breast cancer patients: the evidence does not support avoiding soy. Discuss with your oncologist, but the current research is reassuring. Do not self-restrict based on extrapolated theory.
• Men: there is no evidence that moderate soy intake affects testosterone, sexual function, or fertility.
• Infant soy formula is a separate question with distinct evidence. The above findings do not apply to neonates.

Common misconceptions

• “Soy gives men man-boobs.” No RCT or meta-analysis documents gynaecomastia at typical dietary soy intake. Case reports involve consumption levels no one would reach through normal eating.

• “My doctor told me to avoid soy because it acts like estrogen.” Isoflavones are SERMs, not estrogens. Their net effect in breast tissue may be anti-estrogenic, not estrogenic — the same pharmacological logic that makes tamoxifen useful in ER+ cancer.

• “The studies that show soy is safe are funded by the soy industry.” Meta-analyses pool studies across funding sources. Viscardi et al. (2025) and Hamilton-Reeves et al. (2010) reach the same conclusions regardless of individual study funder.

• “Soy destroys testosterone — that’s why vegan men are weak.” Male testosterone levels are unchanged in every meta-analysis. Strength outcomes depend on total protein intake and training, not soy consumption.

• “Soy causes breast cancer.” Prospective cohort data do not support this. Isoflavone intake may actually compete with endogenous estradiol at ER+ receptors, which is the opposite of the feared mechanism.

The punchline

The soy-and-hormones fear is three things at once: a structural analogy mistaken for a clinical effect, animal data applied without accounting for species differences, and case reports treated as population-level evidence. Human RCTs and meta-analyses consistently find no meaningful hormonal disruption from soy at realistic dietary intake.

One narrow caveat is real: concurrent iodine deficiency can amplify any thyroid effect. The fix is straightforward iodine adequacy — and that applies to all vegans, soy eaters or not.

For a broader picture of how plant proteins fit together, see the complete protein myth article and the main protein pillar.