Does soy cause breast cancer?

The objection

“Soy is full of isoflavones, plant compounds that act like oestrogen, and many breast cancers are fed by oestrogen. In mice carrying human breast cancer cells, genistein from soy made the tumours grow and even cancelled out tamoxifen. When women with breast cancer were given soy protein before their operations, genes that drive cell growth switched on in some of their tumours. If you have had breast cancer, or it runs in your family, why take the risk?”

The answer

Nobody can randomise thousands of women to eat soy or avoid it for decades, so the best evidence on breast cancer itself comes from recording what healthy women eat and following them for years. The China Kadoorie Biobank study (Wei et al., 2020) did that with 300,852 Chinese women aged 30 to 79, and over an average of 10 years 2,289 of them developed breast cancer. Women in the top quarter of soy intake, with a median of 19.1 mg of isoflavones a day, showed no difference in risk from those in the bottom quarter at 4.5 mg (hazard ratio 1.00, 95% confidence interval 0.81 to 1.22) [1].

When the same team pooled that study with eight other cohorts, 631,498 women and 10,229 cases in all, each extra 10 mg of isoflavones a day was linked to a 3% lower risk (hazard ratio 0.97, 0.95 to 0.99), and the studies largely agreed with one another [1]. A 2023 meta-analysis of seven cohort and 17 case-control studies (Yang et al.) found a similar trend in its cohort studies, where each extra 10 mg a day was linked to a 3% to 7% lower risk depending on the statistical method, and the link did not differ significantly by menopausal status [2]. Cancer Research UK (reviewed 2024) answers the question directly: “No, eating soy does not cause cancer” [4].

Does soy protect against breast cancer?

Soy may protect a little, at intakes few Western women reach. In a 2014 meta-analysis of 35 studies (Chen et al.), Asian women with the highest isoflavone intakes had 41% lower odds of breast cancer both before and after the menopause (odds ratio 0.59 in each group), while Western studies found no clear link [3].

Most of the Asian studies (13 of 17 before the menopause and 13 of 18 after) were case-control studies, which ask women about their diet after diagnosis and are more open to recall bias, and the Asian cohort studies on their own found smaller reductions that were not statistically significant (odds ratios 0.77 and 0.84) [3]. Cohort studies record diet before anyone falls ill, which makes them less subject to recall bias [1].

Yang’s meta-analysis (2023) found the same split: comparing the highest eaters with the lowest, the link was significant in the case-control studies (odds ratio 0.62, 0.50 to 0.76) but not in the cohort studies (0.94, 0.86 to 1.02), and studies whose highest eaters took in less than 10 mg a day found no link at all [2].

The flat Western results may partly reflect how little soy Western women eat, and Wei’s team suggest that a fall of only 3% per 10 mg explains why studies at low and moderate intakes found no significant link [1]. In one pooled study the US women averaged 3.2 mg of isoflavones a day after diagnosis, against 45.9 mg among the Chinese women [10], and in Wei’s review the cohorts that found a lower risk were those whose highest eaters took in around 40 mg a day or more [1]. The American Cancer Society (2025) says there is growing evidence that soy foods may lower breast cancer risk, especially among Asian women [5].

Why do some studies suggest soy feeds tumours?

The suggestion comes mainly from mouse studies, and it rests on the fact that isoflavones bind the body’s oestrogen receptors. They can act like a weak oestrogen, or block the body’s own, more potent one, so the worry is a reasonable one to test [4][5].

In mice with their ovaries removed and human oestrogen-dependent breast cancer cells implanted, dietary genistein made the tumours grow in proportion to the dose, at concentrations the researchers called physiological (Ju et al., 2001) [6]. In the same model it overrode tamoxifen, and the authors urged caution for post-menopausal women on the drug (Ju et al., 2002) [7]. The American Cancer Society points out that mice and rats break down isoflavones differently from people, and that animal studies use much higher doses than studies in people [5].

The human study in the objection is real. Researchers at Memorial Sloan Kettering randomised 140 women with early-stage breast cancer to soy protein powder or a milk-protein placebo, two 25.8 g packets a day for the 7 to 30 days before surgery (Shike et al., 2014) [8]. Soy made no significant difference to how fast tumour cells were dividing or dying, but tumours from women with high blood genistein showed more activity in a set of genes linked to cell growth, including FGFR2 [8]. The authors called this a concern for a subset of women, and noted that the study was short, used a large daily dose of soy, and did not examine what the gene changes mean clinically [8].

Longer trials are mostly reassuring: eight randomised trials in 1,287 women, lasting six months to three years, found that isoflavone-rich foods or supplements did not change breast density, a marker of breast cancer risk, after the menopause, though before it they may raise it slightly (a mean difference of 1.83%, which did not hold in one of three sensitivity analyses) (Hooper et al., 2010) [9].

Isoflavone capsules have also been tested in healthy Western women at high risk of breast cancer. In a six-month double-blind trial with 98 women evaluable (Khan et al., 2012), the capsules did not reduce breast cell proliferation, and among premenopausal women taking them the proliferation marker Ki-67 rose from 1.71 to 2.18 [16]. That rise did not differ significantly from the change on placebo, but the authors called it a possible adverse effect in premenopausal women [16].

Is soy safe after breast cancer?

Soy foods appear safe after breast cancer: studies of survivors point to no harm and possibly some benefit. A pooled analysis of 9,514 women from two US cohorts and one Chinese cohort, diagnosed between 1991 and 2006 and followed for an average of 7.4 years, recorded 1,348 recurrences and 1,171 deaths (Nechuta et al., 2012) [10].

Women eating 10 mg or more of isoflavones a day, a level reached by only about a tenth of the US women but by nine in ten of the Chinese women, had a 25% lower risk of recurrence than those eating under 4 mg (hazard ratio 0.75, 0.61 to 0.92), and fewer deaths by margins that were not statistically significant (0.83 for death from breast cancer and 0.87 from any cause) [10]. A meta-analysis of five cohorts and 11,206 patients linked soy after diagnosis to lower mortality (Chi et al., 2013) [11].

Analysis	Recurrence	Death
Nechuta et al. (2012), 9,514 survivors [10]	0.75 (0.61 to 0.92)	0.87 (0.70 to 1.10)
Chi et al. (2013), 11,206 patients [11]	0.74 (0.64 to 0.85)	0.84 (0.71 to 0.99)
Qiu and Jiang (2019), 12 articles [12]	Not given in abstract	0.80 (0.62 to 1.04)
Song et al. (2024), 14 studies and a new cohort [13]	0.81 (0.67 to 0.98)	0.85 (0.76 to 0.96)

The figures are hazard ratios with 95% confidence intervals for the highest soy or isoflavone intakes against the lowest (in Nechuta’s study, 10 mg of isoflavones a day or more against under 4 mg), and the first three measured intake after diagnosis. Below 1.0 means fewer events among women eating more soy, and an interval that crosses 1.0 is not statistically significant. Chi et al. report mortality without naming the cause; the other death figures are for any cause. These analyses overlap, so they are not independent confirmations: Qiu and Jiang’s figure for death from breast cancer (0.83, 0.64 to 1.07) is identical to Nechuta’s [10][12].

Two of these analyses add caveats. In Qiu and Jiang’s review, soy eaten after diagnosis was linked to fewer recurrences but not significantly to survival [12], and Song’s own cohort of 592 Korean survivors, with 47 recurrences, found no clear link either way (hazard ratio 1.29, 0.60 to 2.78) [13].

An independent expert panel for the World Cancer Research Fund’s Global Cancer Update Programme (CUP Global, 2023) graded the evidence that isoflavones from soy foods lower recurrence and death after breast cancer as limited-suggestive, a grade below the “probable” it gave to the link between body fat after diagnosis and death [18][19]. Its figures for recurrence and for death from breast cancer are those of Nechuta’s pooled analysis [19].

All of these are observational studies of what women chose to eat, and they cannot say whether starting soy, or eating much more of it, after diagnosis has the same effect. A 2024 review of 32 studies (van Die et al.) linked soy isoflavones to a 26% lower risk of recurrence (hazard ratio 0.74, 0.60 to 0.92), with the greatest fall at 60 mg a day and no further fall above that, and its authors said evidence is still needed on starting soy or eating substantially more of it after diagnosis [20]. The Chinese survivors in Nechuta’s study averaged 45.9 mg a day [10].

The fear about oestrogen-sensitive tumours and tamoxifen has been tested in people too. Among the 6,690 women in Nechuta’s study whose tumours were oestrogen-receptor positive, eating 10 mg or more a day was not linked to more recurrences (hazard ratio 0.81, 0.63 to 1.04). Women taking tamoxifen who ate the most soy had a 37% lower risk of recurrence than women taking neither tamoxifen nor much soy (0.63, 0.46 to 0.87), against 0.78 (0.62 to 0.99) for tamoxifen users who ate little soy, so there was no sign that soy blunted the drug [10]. Nor does the study show that soy boosts tamoxifen, since its test of whether soy’s link with recurrence differed between women who did and did not take the drug was far from statistically significant (P = 0.99) [10]. The authors could not rule out that soy eaters lived more healthily in ways the study did not measure, and studies like these cannot prove cause [10].

Tamoxifen is not the only hormone treatment, and for aromatase inhibitors such as anastrozole the human evidence is thinner. In a cohort of 524 Chinese women on hormone therapy after surgery (Kang et al., 2010), post-menopausal women eating more than 42.3 mg of isoflavones a day had a 33% lower risk of recurrence than those eating under 15.2 mg (hazard ratio 0.67, 0.54 to 0.85), and among the 86 women taking anastrozole the figure was similar (0.65, 0.47 to 0.85) [17]. The authors said large multicentre trials and further observational studies are needed [17].

Supplements are a separate question. The American Cancer Society (2025) calls soy foods healthy and safe, but does not recommend soy supplements for cancer prevention, because they can contain much higher amounts of isoflavones than food [5]. The European Food Safety Authority (2015) found no evidence that isoflavones at the levels typical of food supplements harm the breast, uterus or thyroid in post-menopausal women, but said there were no data to judge their safety for women with a personal or family history of cancer [14].

If you have had breast cancer, or it runs in your family, keep to foods such as tofu, soya milk and edamame, and ask your oncology team or GP before taking any isoflavone supplement. If you are having treatment for breast cancer, speak to your oncology team or GP before changing or stopping tamoxifen, an aromatase inhibitor or any other medicine.

Does soy affect prostate cancer risk?

Soy has not been linked to a higher risk of prostate cancer, and may be linked to a lower one. A 2018 meta-analysis of 30 studies and 266,699 participants (Applegate et al.) linked eating more soy foods to a 29% lower risk of prostate cancer (relative risk 0.71, 0.58 to 0.85), though the studies based on cohorts, which are less prone to bias than case-control studies, found a smaller 10% reduction (0.90, 0.82 to 0.99) [15].

Total isoflavone intake showed no link (1.03, 0.97 to 1.09), the few studies of advanced prostate cancer found no association, and the authors advise caution, since other differences in diet and lifestyle could contribute and studies that rely on recall are open to recall bias [15]. Whether soy changes men’s hormones is covered in the answer on soy and men.