Estradiol levels for bone health: what the science actually says

What did @drmaryclaire actually say?

The claim is specific: there is a threshold estradiol level around 60-70 pg/mL that produces maximum bone density benefit, and going higher actually causes a dip in lumbar bone density. She references a 2021 study of over 800 postmenopausal patients measuring serum estradiol against lumbar spine BMD. She also draws a useful distinction between stopping bone loss versus actively building new bone, suggesting those require different estradiol levels.

She also says "hormone therapy at almost any commercially available dose is protective against osteoporotic fracture," which is a meaningful baseline claim before she gets into the optimization question. The framing is clinical and directed at patients already on or considering hormone therapy.

Does the science back this up?

Partly, yes. The relationship between endogenous estradiol levels and bone mineral density in postmenopausal women is well-documented. The 2021 study she appears to reference aligns with work published around that time, and the general finding that BMD tracks with estradiol in a non-linear way has some support in the literature.

Cauley et al. (2010, Journal of Bone and Mineral Research) established that postmenopausal women with estradiol levels below 5 pg/mL had significantly higher fracture risk, and that benefit increased with higher levels, but with diminishing returns. Santen et al. (2017, Endocrine Reviews) also noted that breast and bone tissue responses to estradiol show different dose-response curves. The idea of a "ceiling" around 70 pg/mL for bone density is plausible, but the dip she describes above that threshold is less firmly established and relies heavily on the single 2021 study she cites, which she acknowledges has not entered clinical guidelines.

What did they get right, and what's missing?

The osteoclast-osteoblast explanation is accurate and clearly communicated. The Pac-Man analogy is reductive but not wrong. Her framing that "any estrogen will help" aligns with evidence that even low-dose transdermal estradiol reduces bone resorption markers.

What's missing is important context. First, the 2021 study she references, which appears to be observational, cannot establish causation. Women with higher endogenous estradiol may differ in body composition, metabolic health, and aromatization capacity in ways that confound the BMD association. Second, the "dip" above 70 pg/mL is a correlation finding in a cross-sectional analysis, not a signal that higher estradiol damages bone. Third, she does not mention that the relationship between exogenous estradiol doses and serum levels varies significantly by route of administration, body weight, and individual metabolism. A serum level of 70 pg/mL means very different things depending on how someone got there. Fourth, bone density is a surrogate endpoint. Fracture reduction is what matters clinically, and the evidence base for fracture reduction at specific serum levels is thinner than the BMD data.

What should you actually know?

The 60-70 pg/mL target she describes is not in any major clinical guideline, including those from the Menopause Society or the Endocrine Society, and she correctly acknowledges this. That does not make it wrong, but it means it is being used as an off-label clinical heuristic, not an evidence-based standard of care.

The Menopause Society's 2022 position statement on hormone therapy notes that individualized therapy based on symptoms and risk factors is appropriate, and that routine monitoring of serum hormone levels is not standard practice for most patients. Some clinicians do use serum estradiol to guide therapy, particularly for bone health optimization, but this is provider-dependent. Patients who have osteopenia or osteoporosis and are considering hormone therapy should have a dedicated conversation with a clinician about DEXA scan monitoring, fracture risk calculators like FRAX, and whether hormone therapy is appropriate alongside other bone-protective agents. A single serum estradiol number should not be the whole clinical picture.