Is soy good or bad for your bones after menopause?
Soy isoflavones probably do not protect the bones of postmenopausal Western women in any clinically meaningful way. It is better to focus on proven measures such as adequate calcium, vitamin D and exercise.
Soy contains isoflavones, compounds that bear some resemblance to the female hormone oestrogen. That sounds promising for bone health after menopause, but the real-world effect in women is disappointing. A meta-analysis of twelve small studies in postmenopausal Western women found no statistically significant effect of soy isoflavone supplements on lumbar spine bone density. Higher doses or longer treatment made no difference in that analysis.
In Asian women following a traditionally soy-rich diet, epidemiological studies do show signs of better bone protection. Whether that is due to soy or to other differences in lifestyle and dietary patterns cannot be determined. Clinical trials in Western women contradict one another, and multiple reviewers conclude that the available data do not consistently support a protective effect for Western women.
One small study (60 women, 12 months) did find a significant improvement in bone density and bone turnover markers, but it combined equol (a compound some people produce after eating fermented soy) with resveratrol. Which of the two made the difference cannot be derived from that study.
In laboratory and animal research there are indications that isoflavones can stimulate bone-building cells and inhibit bone-breaking cells. That mechanism is biologically plausible, but so far it does not translate consistently into measurable benefits in humans. Whether soy isoflavones work synergistically with vitamin D, as has been described theoretically, has never been tested in a clinical trial.
Soy as part of a varied, healthy diet, alongside adequate calcium, vitamin D and exercise, is probably not harmful. But if you want to actively protect your bones after menopause, soy as a supplement or functional food is too uncertain to rely on.
Eight sources: one small RCT (n=60), one meta-analysis of 12 RCTs (n=1433), multiple systematic reviews and early overview articles. The meta-analysis is the strongest source for the main question; most other sources are reviews of observational or mixed evidence.