GLP-1 medications and bone density: what a new study found — and what it doesn't prove

• A Weill Cornell retrospective cohort study (n=255 GLP-1 users vs. 255 matched controls, PubMed 41655226) found greater annualized hip bone loss in non-diabetic GLP-1 users: -1.0% vs. -0.6% in controls (p=0.04)
• Both groups lost bone density; the GLP-1 group lost more in the non-diabetic subgroup
• No significant between-group difference was seen in patients with diabetes
• Study authors suggest weight loss itself — not a direct drug effect — may be driving the bone changes in non-diabetic patients
• This is observational, retrospective data from a single center; it cannot establish causation
• Separate data presented at ENDO 2026 linked semaglutide to a 15% lower fracture risk compared to other weight-loss medications — evidence of the complexity here

The study, published in the Journal of Clinical Endocrinology & Metabolism (PubMed 41655226, February 8, 2026), was a single-center retrospective cohort at Weill Cornell Medical College in New York.

Researchers included 255 patients using semaglutide or tirzepatide for at least 6 months, matched by age, sex, BMI, and diabetes status to 255 non-users who had at least two DXA (bone density) scans over the same period. The median follow-up was 17 months. Participants were 92% female, with a mean age of 64 years.

The primary finding: in patients without diabetes, GLP-1 users had greater annualized total hip (TH) bone loss than controls — -1.0% vs. -0.6% (p=0.04). That's a difference of 0.4 percentage points in annualized hip bone density change.

Importantly, both groups lost bone density. This wasn't a comparison of loss vs. gain — it was a comparison of how much each group lost over the same period. GLP-1 users without diabetes lost somewhat more.

In patients with diabetes, total hip bone loss was similar between the GLP-1 group and controls. No statistically significant difference was observed in that subgroup.

The divergence between diabetic and non-diabetic patients is one of the study's more interesting findings — and it points toward an explanation.

People with type 2 diabetes generally experience less weight loss from GLP-1 medications than people without diabetes. The study found that weight loss in the GLP-1 group was directly correlated with hip bone loss (r=0.32, p<0.01). In other words, greater weight loss tracked with greater bone loss.

If weight loss is the main driver, then the reason the non-diabetic group showed a bigger gap makes sense: they lost more weight, and with that came more bone loss. The diabetic group's smaller weight loss may explain why their bone changes were similar to controls.

The study authors concluded that weight loss — not a direct pharmacological effect of semaglutide or tirzepatide on bone — may be the primary driver of bone changes in non-diabetic patients.

This is consistent with what's known about weight-loss-related bone changes more broadly. Caloric restriction and bariatric surgery are both associated with bone density loss, likely because the mechanical unloading of bones that comes with losing body mass reduces the stimulus for bone remodeling. GLP-1 drugs may be producing a similar effect through weight loss rather than a specific action on bone tissue.

This distinction matters clinically: if weight loss is the mechanism, then the conversation about bone health applies to anyone losing significant weight through any effective method — not just GLP-1 users specifically.

This is an observational, retrospective, single-center study. It has several important limitations:

• It cannot establish causation. The association between GLP-1 use and greater bone loss doesn't prove the drug caused it — weight loss is a confounder that the authors themselves identify as the likely driver.
• It's a single center. Weill Cornell patient populations may differ from other settings in ways that affect the findings.
• Fracture risk wasn't the outcome measured. Bone density change is a surrogate measure; whether the observed change translates to increased fracture risk is a separate question.
• Median follow-up was 17 months. Longer-term bone effects aren't captured here.

Furthermore, data presented at the ENDO 2026 conference — from a different study — found that semaglutide use was associated with a 15% lower fracture risk compared to alternative weight-loss medications. This points to the complexity of the GLP-1/bone relationship and suggests that BMD reduction doesn't automatically translate to increased fractures in this context.

• Whether the bone density differences observed in this study translate to increased fracture risk over time
• Whether the effects differ between semaglutide and tirzepatide (the study combined them)
• What the long-term (>2 year) bone effects of GLP-1 therapy look like in people without diabetes
• Whether co-interventions — resistance training, protein intake, calcium and vitamin D — can meaningfully offset weight-loss-related bone changes in GLP-1 users

• Do I have any baseline risk factors for bone loss or fracture that make bone monitoring especially important for me?
• Should I have a DXA scan before or during GLP-1 therapy, and if so, how often?
• Are there adjustments to my diet, exercise routine, or supplements that make sense given my GLP-1 use and weight loss?
• What does the overall evidence on fracture risk — not just bone density — suggest for someone with my profile?

• DXA scan results (total hip, lumbar spine, and femoral neck) if monitoring is part of your care
• Weight loss trajectory alongside any bone monitoring results
• Calcium and vitamin D intake as part of your overall nutritional picture

One study's findings don't change clinical practice guidelines. But they do give you a specific, evidence-grounded question to bring to your clinician: given the weight I'm losing on this medication, is my bone health something we should be monitoring?

Medical disclaimer: This content is for educational purposes only and is not medical advice. Always consult a licensed healthcare professional before starting, stopping, or changing any treatment.

Sources

• PubMed 41655226 — Skeletal Effect of Semaglutide and Tirzepatide in Patients with Increased Risk of Fractures (J Clin Endocrinol Metab, February 2026) — pubmed.ncbi.nlm.nih.gov/41655226
• eCommons Cornell institutional page — ecommons.cornell.edu/entities/publication/a6d749be-ae8d-4242-83c8-b2bc70393add
• ENDO 2026 fracture risk signal — endocrine.org/news-and-advocacy/news-room/2026/jairo-norena-press-release-endo-2026
• Endocrinology Advisor coverage — endocrinologyadvisor.com/news/glp1ra-bone-mineral-density-loss-diabetes-status