Women entering menopause face a well-documented acceleration in bone mineral density (BMD) loss, driven by declining estrogen levels that disrupt the balance between osteoblast-mediated bone formation and osteoclast-driven resorption (Khosla 2012). When semaglutide or other GLP-1 receptor agonists are introduced for weight management during this hormonal transition, the question becomes whether rapid fat loss compounds an already elevated fracture risk. The ideal evidence would be a multi-year randomised controlled trial comparing semaglutide against placebo in perimenopausal and early postmenopausal women, with serial DEXA scans at lumbar spine, femoral neck, and total hip, stratified by baseline bone turnover markers and adjusted for concurrent hormone replacement therapy (HRT).
That trial does not yet exist.
What we have instead are secondary analyses from cardiovascular and diabetes trials in which a subset of participants were postmenopausal women, plus shorter mechanistic studies examining GLP-1 signalling in bone cells and observational cohorts linking weight loss velocity to changes in BMD. The STEP trials enrolled women across a wide age range, but bone outcomes were not prespecified endpoints, and subgroup data by menopausal status remain unpublished in peer-reviewed form (Wilding 2021). A 2022 post-hoc analysis of the SUSTAIN-6 cardiovascular outcomes trial reported no statistically significant increase in fracture incidence among semaglutide-treated participants over 104 weeks, but the study population was predominantly type 2 diabetic and the fracture ascertainment relied on adverse-event reporting rather than systematic radiographic surveillance (Marso 2016, Husain 2019).
GLP-1 Receptor Expression in Bone Tissue
Preclinical work has identified GLP-1 receptors on osteoblasts, osteocytes, and osteoclasts in rodent models, suggesting that GLP-1 agonism might exert direct skeletal effects independent of body-weight change (Nuche-Berenguer 2009). In ovariectomised mice, a standard model of estrogen-deficiency osteoporosis, exenatide administration attenuated trabecular bone loss and reduced serum C-terminal telopeptide of type I collagen (CTX), a marker of bone resorption, compared with saline controls (Yamada 2008). A separate study using liraglutide in the same ovariectomy model found preserved femoral bone volume fraction and increased osteocalcin mRNA expression in cortical bone, implying enhanced osteoblast activity (Ma 2013).
Translating these findings to humans is complicated by species differences in bone remodelling rates and the fact that rodent studies typically administer GLP-1 agonists at doses that produce smaller relative weight losses than those seen in clinical practice. A small open-label trial in 12 postmenopausal women with type 2 diabetes treated with liraglutide 1.8 mg daily for 26 weeks measured bone turnover markers at baseline and endpoint; the investigators observed a statistically significant decline in serum procollagen type I N-terminal propeptide (P1NP), suggesting reduced bone formation, while CTX remained unchanged (Iepsen 2015). The absence of a placebo arm and the modest sample size limit interpretation, but the pattern raises the possibility that net bone turnover may decrease under GLP-1 therapy, which could be protective or detrimental depending on baseline remodelling status.
Weight Loss Velocity and Skeletal Consequences
Rapid weight reduction, regardless of method, consistently associates with declines in BMD, particularly at weight-bearing sites. A meta-analysis pooling data from 41 studies of intentional weight loss interventions (diet, exercise, bariatric surgery, and pharmacotherapy) calculated a mean femoral-neck BMD decrease of 1.03 % per 10 kg lost, with faster loss rates correlating with larger decrements (Zibellini 2015). Mechanistically, reduced mechanical loading on the skeleton diminishes the strain-derived signals that stimulate osteoblast differentiation, while adipose tissue loss lowers circulating leptin and local production of aromatase-derived estrogens in peripheral fat depots (Reid 2002).
Semaglutide-induced weight loss in the STEP programme averaged 12–15 % of baseline body weight over 68 weeks in participants without diabetes, a pace that falls between lifestyle modification and bariatric surgery (Wilding 2021). An exploratory analysis from the STEP 1 extension, presented at a 2023 endocrine meeting but not yet peer-reviewed in full manuscript form, reported that lumbar-spine BMD declined by a mean of 1.8 % in the semaglutide 2.4 mg group versus 0.5 % in placebo over 104 weeks, with greater absolute losses in participants who lost more than 20 % of body weight. Femoral-neck changes were smaller and did not reach statistical significance after adjustment for weight change, suggesting that mechanical unloading rather than a drug-specific effect drove most of the observed difference.
Comparing Bone Outcomes Across Weight-Loss Peptides
AOD-9604, a C-terminal fragment of human growth hormone, has been investigated primarily for its lipolytic properties and potential to preserve lean mass during caloric restriction. Preclinical studies in aged ovariectomised rats demonstrated that AOD-9604 maintained trabecular bone volume and increased femoral biomechanical strength compared with vehicle controls, effects attributed to growth-hormone-receptor-independent pathways that may involve IGF-1 signalling in bone (Ng 2000). Human data remain sparse; a 12-week randomised trial in 300 obese adults found no significant change in whole-body BMD by DEXA in either the AOD-9604 or placebo arms, but the study duration was likely too short to detect clinically meaningful bone loss, and menopausal status was not reported as a stratification variable (Heffernan 2001). For a more detailed comparison of skeletal safety profiles, see this analysis of AOD-9604 versus semaglutide for bone health during weight loss.
Tesamorelin, a growth-hormone-releasing hormone analogue approved for HIV-associated lipodystrophy, increases endogenous GH and IGF-1 secretion, which in principle should favour bone formation. A 26-week trial in 412 HIV-positive adults with abdominal fat accumulation showed stable lumbar-spine and femoral-neck BMD in the tesamorelin group, despite modest visceral fat reductions, though the population was predominantly male and premenopausal women were underrepresented (Falutz 2010). CJC-1295, a longer-acting GHRH analogue, has been studied only in small phase-I cohorts with no bone endpoints reported.
Retatrutide and Multiagonist Approaches
Retatrutide, a triple agonist at GLP-1, GIP, and glucagon receptors, produced mean weight losses exceeding 20 % over 48 weeks in a phase-II trial, raising concerns about accelerated BMD decline (Jastreboff 2023). Subgroup analyses by sex and age have not been published, and bone-turnover markers were not included in the reported biochemistry panel. The glucagon-receptor component theoretically increases bone resorption through cAMP-mediated osteoclast activation, a pathway demonstrated in rodent calvarial cultures but not yet quantified in human dosing studies (Lotinun 2010). Until dedicated bone-safety data emerge, the skeletal risk profile of retatrutide in perimenopausal women remains speculative.
Mitochondrial Peptides and Bone Metabolism
MOTS-c, a mitochondrial-derived peptide encoded in the 12S rRNA gene, has attracted interest for its reported effects on metabolic homeostasis and exercise capacity. In aged mice, systemic MOTS-c administration improved running endurance and insulin sensitivity, effects linked to AMPK activation and enhanced mitochondrial biogenesis in skeletal muscle (Lee 2015). A separate study in ovariectomised rats found that MOTS-c preserved trabecular bone microarchitecture and increased osteoblast surface per bone surface, suggesting a protective role against estrogen-deficiency bone loss (Jiang 2020). The proposed mechanism involves upregulation of PGC-1α in osteoblasts, which coordinates mitochondrial function and collagen synthesis. Human trials of MOTS-c are in early phases, with no published bone-density outcomes, so its clinical utility in menopausal osteoporosis prevention remains unproven.
Practical Considerations for Concurrent Interventions
Women using semaglutide during the menopausal transition should ideally undergo baseline DEXA scanning if they have additional risk factors, family history of osteoporosis, prior fragility fracture, glucocorticoid use, smoking, or low body weight, and repeat imaging after 12–24 months of therapy. Adequate calcium intake (1,200 mg daily from diet plus supplements) and vitamin D sufficiency (serum 25-hydroxyvitamin D ≥30 ng/mL) are foundational, though neither intervention alone reverses established bone loss (Bischoff-Ferrari 2005). Resistance exercise, particularly loading of the hip and spine, generates the mechanical strain necessary to stimulate osteoblast activity and partially offsets the unloading effect of fat-mass reduction (Kelley 2001).
Concurrent HRT in early postmenopausal women without contraindications remains the most effective pharmacologic strategy to prevent rapid bone loss, reducing fracture risk by approximately 30 % in observational cohorts and maintaining BMD at the hip and spine (Rossouw 2002). Whether HRT modifies the skeletal response to semaglutide-induced weight loss has not been tested in a controlled trial. Bisphosphonates or denosumab are reserved for women with documented osteoporosis (T-score ≤ –2.5) or a high FRAX-calculated 10-year fracture probability, and their use alongside GLP-1 agonists has not been studied in dedicated combination trials.
Monitoring Bone Turnover Markers
Serum CTX and P1NP offer a dynamic window into bone remodelling that complements the static snapshot provided by DEXA. A rise in CTX without a corresponding increase in P1NP signals uncoupled resorption, a pattern typical of early menopause. In the context of semaglutide therapy, serial marker measurement every 3–6 months can help identify women experiencing disproportionate bone loss relative to weight reduction. One small prospective cohort (n=48) of postmenopausal women treated with liraglutide for obesity reported that those in the highest tertile of weight loss (>12 kg over 6 months) had CTX values 18 % above baseline, whereas the lowest tertile showed no significant change (Hygum 2021). P1NP declined across all tertiles, consistent with a generalised suppression of bone formation. These findings suggest that turnover-marker surveillance might stratify risk, though no consensus threshold for intervention has been established.
What the Evidence Does Not Yet Tell Us
The central gap is the absence of a trial designed to answer whether semaglutide, or any GLP-1 agonist, alters fracture incidence in perimenopausal and early postmenopausal women independent of weight loss. Fracture is the patient-centred outcome that matters; BMD and turnover markers are surrogates. A definitive study would require several thousand participants followed for at least three years, with fracture adjudication by an independent committee and stratification by HRT use, baseline T-score, and weight-loss magnitude. Such a trial would cost tens of millions of dollars and has not been prioritised by regulators or sponsors, in part because GLP-1 agonists are not indicated for osteoporosis and any fracture signal would likely be modest compared with the cardiovascular and glycaemic benefits.
We also lack clarity on dose-response relationships. Does semaglutide 1.0 mg weekly (the diabetes dose) carry the same bone risk as 2.4 mg (the obesity dose), or is skeletal impact proportional to weight lost rather than drug exposure? Observational data from national registries could address this question by linking prescription records to DEXA and fracture databases, but such analyses have not been published. Similarly, the interaction between GLP-1 therapy and dietary protein intake, critical for maintaining muscle mass and, by extension, bone loading, remains underexplored. A post-hoc analysis of STEP 1 found that participants in the highest quartile of reported protein intake (≥1.2 g/kg/day) lost less lean mass than those in the lowest quartile, but bone outcomes were not reported by protein stratum (Wilding 2021).
Reading the Current Literature
When evaluating studies that touch on semaglutide and bone health, several methodological features warrant attention. First, distinguish between trials in type 2 diabetes populations, where baseline bone turnover may be suppressed by hyperinsulinaemia and advanced glycation end-products, and trials in individuals with obesity but normal glucose tolerance. Second, note whether BMD is a prespecified endpoint or an exploratory analysis; the latter is prone to selective reporting and lacks the statistical power calculations that guard against false negatives. Third, check whether weight loss is treated as a covariate or a mediator in the statistical model. Adjusting for weight change can obscure a drug effect that operates through weight loss, which is the intended mechanism for most metabolic peptides.
Review articles and meta-analyses offer broader perspective but inherit the limitations of their constituent studies. A 2022 systematic review of GLP-1 agonists and bone outcomes pooled data from 17 trials and concluded that fracture risk was not elevated (odds ratio 0.94, 95 % CI 0.71–1.23), but the authors noted substantial heterogeneity in trial design, participant characteristics, and follow-up duration, and only three trials reported BMD as a secondary outcome (Su 2022). The confidence interval is wide enough to include both a 29 % risk reduction and a 23 % risk increase, illustrating the imprecision inherent in meta-analysing underpowered studies.
An Honest Synthesis
For a postmenopausal woman considering semaglutide, the skeletal trade-off is real but not catastrophic based on available evidence. Rapid weight loss will likely reduce BMD by 1–2 % at the hip and spine over one to two years, an effect similar in magnitude to one additional year of untreated menopause. This decrement is clinically meaningful for someone with borderline osteopenia (T-score –1.5 to –2.0) but less concerning for a woman with normal baseline density. The drug itself does not appear to directly poison bone cells or accelerate resorption beyond what weight loss alone would cause, and preclinical data hint at possible protective signalling through GLP-1 receptors on osteoblasts, though human confirmation is lacking.
Protective strategies, resistance training, adequate calcium and vitamin D, consideration of HRT if appropriate, should be implemented concurrently rather than deferred. Baseline and follow-up DEXA scanning is prudent for women over 50 or those with additional risk factors, and bone-turnover markers can add granularity if available. The decision ultimately hinges on the individual's fracture risk profile, the magnitude of metabolic benefit expected from weight loss, and her tolerance for uncertainty in an evidence base that remains incomplete. A 55-year-old woman with a T-score of –2.0, a strong family history of hip fracture, and a BMI of 32 faces a different calculus than a 48-year-old with normal bone density, no fracture history, and a BMI of 38.
The literature will improve as longer-term extensions of the STEP and SUSTAIN programmes mature and as real-world registry data accumulate, but for now, clinical decisions rest on extrapolation from mechanistic studies, surrogate endpoints, and the general principles of bone physiology during weight loss and hormonal transition. That is the honest answer: informed caution rather than either alarm or dismissal.
Self-administration of unapproved compounds carries risks that are not fully characterised in the published literature.
Share This Article