Did the STEP 1 and SURMOUNT-1 trials measure bone density?

No. Both trials ran DXA sub-studies, and both were designed to demonstrate that the weight lost was fat. STEP 1's registered DXA outcomes are total fat mass, lean body mass and visceral fat mass; SURMOUNT-1's DXA addendum had percent change in total body fat mass as its primary objective. The words bone, BMD, mineral density, osteoporosis and fracture appear zero times in either trial's protocol or statistical analysis plan as filed with ClinicalTrials.gov.

Is bone loss on a GLP-1 caused by the drug or by the weight loss?

Bone density falls with any substantial weight loss, including diet-induced and surgical. In a randomised trial of 160 obese older adults, every exercise group lost 9 percent of body weight on the same weight-management program, and over six months total hip bone density fell 3 percent with aerobic training, 1 percent with combined training and 0.5 percent with resistance training. Bariatric surgery is associated with a 21 to 44 percent higher risk of all fractures in observational studies. A retrospective study of 255 adults on semaglutide or tirzepatide, 92 percent female with a mean age of 64, found weight loss correlated with bone loss within the drug group, and its authors concluded that weight loss appears to drive bone loss in patients without diabetes.

What is the strongest randomised evidence on GLP-1s and bone density?

A predefined secondary analysis of a Danish randomised trial published in JAMA Network Open on June 3, 2024. After an 8-week low-calorie diet, 195 adults with obesity were randomised for 52 weeks to exercise, liraglutide 3.0 mg, both, or placebo, with site-specific bone mineral density as the primary outcome. Total hip bone density fell 0.026 g/cm2 on liraglutide alone versus 0.013 with exercise, 0.019 with the combination and 0.012 with placebo. Against placebo, liraglutide alone was 0.013 g/cm2 lower at the hip and 0.019 lower at the lumbar spine; the combination arm did not differ from placebo at either site. The drug was liraglutide, not semaglutide or tirzepatide, mean age was 42.8 years, and only 25 of the 195 participants were recorded as menopausal.

Should a postmenopausal woman get a DXA scan before starting a GLP-1?

No guideline says so, and nothing on this page implies anyone needs a scan. The US Preventive Services Task Force recommendation of January 14, 2025 recommends screening for osteoporosis in women 65 years or older, and in postmenopausal women younger than 65 who are at increased risk for an osteoporotic fracture as estimated by clinical risk assessment. The Bone Health and Osteoporosis Foundation's 2022 Clinician's Guide lists younger postmenopausal women and women in the menopausal transition with clinical risk factors for fracture among those in whom bone density testing should be considered. Neither document mentions GLP-1 drugs. Whether an individual meets those thresholds is a clinical judgment about her own risk factors.

Women's Health

GLP-1s and bone: what is measured, what is inferred, and what postmenopausal women should ask for

Q: Does Ozempic or Wegovy cause bone loss?

No pivotal obesity trial of semaglutide measured bone mineral density: STEP 1 scanned a DXA sub-population for fat and lean mass only. The current Ozempic label contains no fracture, osteoporosis or bone-density language at all. The current Wegovy label does: in an Adverse Reactions subsection headed Fractures, it reports that in its cardiovascular outcomes trial more fractures of the hip and pelvis were reported on Wegovy than on placebo in female patients, 1 percent (24 of 2,448) versus 0.2 percent (5 of 2,424), and in patients aged 75 and older, 2.4 percent (17 of 703) versus 0.6 percent (4 of 663). Those are adverse-event counts from a trial designed to test cardiovascular outcomes over a median of about 37 months in adults with established cardiovascular disease, not a fracture endpoint, and fracture does not appear in Warnings and Precautions.

By the US Health Digest editorial team · Published August 30, 2026 · Every claim linked to its primary source

Neither STEP 1 nor SURMOUNT-1 measured bone mineral density. Both ran DXA sub-studies, and both were designed to show that the weight lost was fat — the trial protocols and statistical analysis plans contain no bone endpoint at all. The strongest randomised evidence on bone comes from a different drug in a younger population: in a 195-person Danish trial, liraglutide alone lowered hip and spine density more than exercise alone. One US label does print a fracture imbalance in women — Wegovy's, from its cardiovascular outcomes trial. The current Zepbound, Ozempic and Mounjaro labels do not use the word "fracture" anywhere.

What the pivotal obesity trials actually measured

Both flagship trials scanned people. Neither scanned their bones.

STEP 1 (semaglutide 2.4 mg, 14.9% weight loss at 68 weeks, PMID 33567185) ran a DXA sub-population of approximately 135 randomised participants. Its registered outcomes from those scans are total fat mass, lean body mass and visceral fat mass — nothing else. The protocol states the purpose plainly: the sub-population was scanned "to demonstrate that weight loss is primarily caused by reduction in fat mass, in accordance with the EMA and FDA guidelines." We searched the full protocol and statistical analysis plan: "bone", "BMD", "mineral density", "osteoporosis" and "fracture" appear zero times in either, while "DEXA" appears 17 and 14 times.

SURMOUNT-1 (tirzepatide, up to 20.9% at 72 weeks, PMID 35658024) registered no body-composition outcome at all. Its statistical analysis plan describes a separate DXA addendum whose "primary objective… is to demonstrate that tirzepatide (5 mg, 10 mg, and 15 mg combined) QW is superior to placebo in percent change in total body fat mass from baseline to 72 weeks," with lean mass secondary. The word "bone" does not appear in that plan, nor in the SURMOUNT-1 protocol.

No supplement can close that gap: a protocol that never specified bone and a plan that never analysed it leave nothing to find. What the sub-studies do say about muscle is in GLP-1s and lean mass.

The one bone signal printed on a US label

It is on Wegovy, and the comparison it leads with is in female patients. The current label (SPL version 19, effective June 18, 2026) carries a subsection headed "Fractures" inside Adverse Reactions, section 6.1:

"In the CV outcomes trial in adults, more fractures of the hip and pelvis were reported on WEGOVY than on placebo in female patients: 1% (24/2,448) vs. 0.2% (5/2,424), and in patients ages 75 years and older: 2.4% (17/703) vs. 0.6% (4/663), respectively."

Read the fine print. The trial is SELECT: 8,803 patients exposed to semaglutide for a median of 37.3 months against 8,801 on placebo for 38.6 months, in adults with established cardiovascular disease. The registry's baseline table confirms the label's female denominators exactly: 2,448 women on semaglutide and 2,424 on placebo. Only that first pair is sex-restricted — 703 and 663 are the patients aged 75 and over in each arm, men included, and section 8.5 confirms 703 as the count of Wegovy-treated patients in that age band. The same subsection then reports fractures in a separate trial in adults with MASH, and adds that fractures "were reported in both males and females with a median age of 61 years (range, 44 to 75)." These are adverse-event counts from a trial designed to test heart attacks and strokes, and the label states that safety collection there "was limited to serious adverse events (including death), adverse events leading to discontinuation, and adverse events of special interest." Fracture appears nowhere in Warnings and Precautions, which run 5.1 to 5.11 without a bone entry.

The other three labels are silent. We searched the full structured product labels for Zepbound (v38, April 22, 2026), Ozempic (v20, June 1, 2026) and Mounjaro (v38, April 22, 2026) end to end. "Fracture", "osteoporosis" and "bone mineral density" appear zero times in each; the only substantive "bone" hits are the rat bone-marrow micronucleus assay and, on Ozempic, skeletal abnormalities in animal offspring; the rest are the phrase "peptide backbone". The asymmetry may be about trial programmes rather than molecules: semaglutide has a three-year, 17,604-participant outcomes trial reported, while the tirzepatide obesity trials ran 72 weeks.

The randomised evidence on bone density itself

One randomised trial has published a bone analysis whose stated primary outcome is site-specific bone density, and it used liraglutide, not semaglutide or tirzepatide. Jensen and colleagues, JAMA Network Open, June 3, 2024 is a predefined secondary analysis of a Danish trial: 195 adults with obesity, mean age 42.8, 124 women, put through an 8-week 800 kcal/day diet and then randomised for 52 weeks to exercise, liraglutide 3.0 mg daily, both, or placebo. All four arms share one estimand — estimated mean change in BMD from week −8 (before the diet) to week 52, intention-to-treat.

All four bars share one scale and one estimand: estimated mean within-group change, not placebo-subtracted differences. Against placebo, liraglutide alone was −0.013 g/cm² (95% CI −0.024 to −0.002) at the hip and −0.019 (−0.034 to −0.004) at the spine; the combination arm differed from placebo at neither site.

The authors' conclusion, verbatim: "In this randomized clinical trial, the combination of exercise and GLP-1RA (liraglutide) was the most effective weight loss strategy while preserving bone health. Liraglutide treatment alone reduced BMD at clinically relevant sites more than exercise alone despite similar weight loss." Two limits. The drug is liraglutide, an older and less potent agent. And the trial says almost nothing about postmenopausal women: mean age 42.8, with menopause recorded in 25 of the 195 participants.

The only other randomised BMD data we located is a 20-week, 20-person semaglutide pilot in older adults that found no significant between-group difference in whole-body BMD — too small, and only 5.3% weight loss, to settle anything.

A 2025 meta-analysis of 25 studies in type 2 diabetes pooled the opposite direction: no significant fracture effect (RR 0.80, 95% CI 0.47 to 1.36) and higher BMD than control at the lumbar spine (mean difference 0.07 g/cm², 0.06 to 0.09) and total hip (0.06, 0.04 to 0.07). Note the scale before believing it: 0.07 g/cm² is roughly 6% of a typical adult lumbar spine BMD — a gain on the order of what dedicated osteoporosis treatment is expected to produce, which is a great deal to attribute to a diabetes drug. The review itself asks for "further high-quality clinical studies with sufficient follow-up time."

The confound that explains most of this: weight loss itself

Bone falls when weight falls, whatever is moving the weight. In Villareal and colleagues' 160-person randomised trial in obese older adults (NEJM, 2017) — diet-induced weight loss, no drug — body weight fell 9% in every exercise group, while total hip BMD fell 3% with aerobic training, 1% with combined training and 0.5% with resistance training. Those are within-group changes at six months among the 141 of 160 randomised participants who completed. Same weight loss; three different bone outcomes, depending on what participants did with their muscles. After bariatric surgery the effect is larger: the European Calcified Tissue Society reports that, based on observational studies, "bariatric surgery is associated with a 21-44% higher risk of all fractures."

The most postmenopausal-relevant dataset we found points the same way. A retrospective study at Weill Cornell (JCEM, June 2026) matched 255 adults on semaglutide or tirzepatide for at least six months — 92% female, mean age 64 — to 255 non-users with two DXA scans over the same period. Both groups lost total hip and femoral neck bone by similar amounts. Within the drug group, weight loss correlated with bone loss (r = 0.32 at the total hip, P < .01). Only among patients without diabetes did the drug group lose more total hip BMD than controls, and the gap was −1% versus −0.6% (P = .04). The authors' reading: "These findings suggest GLP-1 RA's effects on bone may differ by DM status, with weight loss driving bone loss in patients without DM." Single-centre, retrospective, median weight loss 5%.

Why the large database studies contradict each other

Four recent analyses drawn from the same US electronic-record network reach four different places. They are not pooled here and must not be read against each other: comparators, outcomes, populations and follow-up windows all differ.

Observational analyses of GLP-1 receptor agonists and skeletal outcomes, all using the TriNetX network, as of August 30, 2026. Different comparators and different outcomes; none establishes causation.
Study	Population and comparator	Outcome	Result (95% CI)
Hamad et al., JAMA Netw Open, July 2026 (PMID 42496972)	133,606 matched adults 50–90 with type 2 diabetes; GLP-1 vs DPP-4 inhibitor	Fragility fracture, 3 years	HR 0.79 (0.76–0.83). In a sensitivity analysis stratified by diabetes status: 0.91 (0.88–0.95) with diabetes, 1.13 (1.04–1.23) without
Huang et al., Diabetes Obes Metab, July 2026 (PMID 42010367)	Obesity with and without diabetes; semaglutide vs empagliflozin, glipizide or usual care	Major osteoporotic fracture	0.69 (0.61–0.77) to 0.84 (0.76–0.93) with diabetes; no significant association without diabetes
Constantine et al., JAAOS, June 2026 (PMID 41429015)	33,210 matched pairs, overweight or obese, no diabetes; GLP-1 vs no GLP-1	Any fracture, 1 year	3.05% vs 2.61%; OR 1.19 (1.09–1.31)
Wu et al., Obesity, June 2026 (PMID 42225263)	18,062 matched pairs aged 50+, obesity, no diabetes; GLP-1 vs other obesity medications	Osteoporosis; major osteoporotic fracture	RR 0.48 (0.43–0.52); RR 0.15 (0.11–0.22)

Two patterns survive. Diabetes status keeps flipping the sign: both studies that stratified on it found benefit in diabetes and either nothing or harm without it. And a relative risk of 0.15 for major osteoporotic fracture — an 85% reduction, from a database — is a magnitude no bone drug achieves, and reads better as a signal about who gets prescribed these medicines than as an effect on the skeleton.

What actually helps, and who says so

Nothing here is a reason to start, stop or change a medication, and nobody here is telling you that you need a scan. The interventions with evidence behind them are the ones you would want during any substantial weight loss.

The Bone Health & Osteoporosis Foundation's Clinician's Guide to Prevention and Treatment of Osteoporosis (2022) lists its universal interventions as "adequate intake of calcium, vitamin D, and protein; regular participation in weight-bearing and muscle-strengthening physical activity; cessation of tobacco use; and recognition and treatment of alcohol abuse." It endorses the Institute of Medicine calcium intakes — 1,200 mg a day for women 51 and older — and for people with osteoporosis recommends a multicomponent programme built around progressive resistance training, balance training, back extensor strengthening and impact loading. It is candid about the ceiling: its glossary defines exercise as an intervention "long associated with healthy bones, despite limited evidence for significant beneficial effect on BMD or fracture risk reductions," and its stronger case is fall prevention. On protein, see protein-first eating on a GLP-1 and eating when appetite is suppressed.

That 2022 guide is the current edition, and it predates the question: it mentions no GLP-1 receptor agonist and no pharmacological weight loss, listing obesity and bariatric surgery among conditions associated with bone loss and nothing about the drugs.

On scanning, the US Preventive Services Task Force recommendation of January 14, 2025 is the US reference, and it does not mention weight-loss drugs either. It "recommends screening for osteoporosis to prevent osteoporotic fractures in women 65 years or older" (B recommendation) and "in postmenopausal women younger than 65 years who are at increased risk for an osteoporotic fracture as estimated by clinical risk assessment" (B recommendation). BHOF's own list of people in whom to consider testing includes "Younger postmenopausal women, women in the menopausal transition, and men aged 50 to 69 years with clinical risk factors for fracture." Whether you meet either threshold is a judgment about your risk factors, not about your prescription.

One study is finally asking the question in the right population: NCT07422987, a prospective cohort of 50 women aged 35 to 60 at the University of Kansas Medical Center, lists bone mineral density among its primary outcomes and estrogen's role in BMD change among its secondary ones, with estimated primary completion June 30, 2027. It is observational, not randomised.

Questions for your prescriber

This is evidence reporting, not medical advice.

Given my own risk factors — age, menopausal status, previous fracture, parental hip fracture, smoking, alcohol, steroid use, low body weight — do I meet a guideline threshold for bone density testing, independently of this medication?
If a baseline scan is indicated for me, is it better done before treatment than after, so there is something to compare against?
The pivotal trials did not measure bone. When you tell me what to expect, is that trial data, database data, or clinical judgment?
What is the plan for resistance training and protein intake alongside the weight loss, and who helps me build it?
Should calcium and vitamin D intake be measured rather than assumed?

Absence claims and how we tested them

Three statements above claim something does not exist, each scoped to searches run on August 30, 2026. One: STEP 1 and SURMOUNT-1 did not measure bone mineral density — both registry records pulled from the ClinicalTrials.gov v2 API and all four filed protocols and statistical analysis plans downloaded and searched; "bone", "BMD", "mineral density", "osteoporosis" and "fracture" return zero hits in all four documents. NEJM blocks this system from both supplementary appendices, so the claim deliberately rests on the sponsors' design documents rather than on the papers. Two: the current Zepbound, Ozempic and Mounjaro labels contain no fracture, osteoporosis or bone-density language — each full SPL XML pulled from the DailyMed API and string-searched, zero hits, with version history confirming all four labels we cite are the newest on file. Three: the 2022 BHOF guide does not address GLP-1 drugs — zero hits in its full text for "GLP-1", "glucagon", "semaglutide", "liraglutide" and "anti-obesity".

Sources

Jensen SBK, Sørensen V, Sandsdal RM, et al. Bone Health After Exercise Alone, GLP-1 Receptor Agonist Treatment, or Combination Treatment: A Secondary Analysis of a Randomized Clinical Trial. JAMA Network Open. 2024;7(6):e2416775. PubMed 38916894; open-access full text (PubMed Central). Tables 1 and 2 used for the chart.
DailyMed. WEGOVY (semaglutide) — current prescribing information. National Library of Medicine; SPL version 19, effective June 18, 2026. Sections 5, 6.1 ("Fractures") and 8.5.
DailyMed. ZEPBOUND (tirzepatide), SPL version 38, effective April 22, 2026; OZEMPIC (semaglutide), SPL version 20, effective June 1, 2026; MOUNJARO (tirzepatide), SPL version 38, effective April 22, 2026. National Library of Medicine.
ClinicalTrials.gov registry records and posted results, read August 30, 2026: NCT03548935 (STEP 1), NCT04184622 (SURMOUNT-1), NCT03574597 (SELECT). National Library of Medicine.
STEP 1 trial documents filed with the registry: study protocol (March 29, 2019) and statistical analysis plan (July 3, 2020). SURMOUNT-1: study protocol (October 4, 2021) and statistical analysis plan (September 28, 2021).
Liu Y, Walzer D, Schmitz S, et al. Skeletal effect of semaglutide and tirzepatide in patients with increased risk of fractures. Journal of Clinical Endocrinology & Metabolism. 2026;111(7):1959-1966. PubMed 41655226.
Tan Y, Liu S, Tang Q. Effect of GLP-1 receptor agonists on bone mineral density, bone metabolism markers, and fracture risk in type 2 diabetes: a systematic review and meta-analysis. Acta Diabetologica. 2025;62(5):589-606. PubMed 39985672.
Dinkla L, Beavers KM, Robbins R, et al. Bone mineral density and turnover response to GLP-1 receptor agonists in older adults with overweight/obesity and prediabetes/type 2 diabetes: a 20-week pilot trial post hoc analysis. Frontiers in Aging. 2025;6:1691007. PubMed 41393101.
Hamad CD, Wiener J, Golzar A, et al. Glucagon-Like Peptide-1 Receptor Agonists and Fragility Fracture Risk in Type 2 Diabetes. JAMA Network Open. 2026;9(7):e2625141. PubMed 42496972; full text (PubMed Central).
Huang YN, Tsou MY, Li PH, et al. Associations of Semaglutide With Skeletal Outcomes in People With Obesity, With and Without Type 2 Diabetes: A Target Trial Emulation. Diabetes, Obesity and Metabolism. 2026;28(7):5834-5847. PubMed 42010367; full text (PubMed Central).
Constantine E, Enthoven L, Kahan R, Pflug EM, Lauder A. The Impact of Glucagon-Like Peptide-1 Receptor Agonists on Fracture Risk in Overweight or Obese, Nondiabetic Patients. Journal of the American Academy of Orthopaedic Surgeons. 2026;34(12):e1610-e1619. PubMed 41429015.
Wu JS, Lin LY, Sun JW, et al. Musculoskeletal Outcomes of Glucagon-Like Peptide-1 Receptor Agonists Versus Other Antiobesity Agents in Nondiabetic Adults. Obesity (Silver Spring). 2026. PubMed 42225263.
Villareal DT, Aguirre L, Gurney AB, et al. Aerobic or Resistance Exercise, or Both, in Dieting Obese Older Adults. New England Journal of Medicine. 2017;376(20):1943-1955. PubMed 28514618.
Paccou J, Tsourdi E, Meier C, et al. Bariatric surgery and skeletal health: A narrative review and position statement for management by the European Calcified Tissue Society (ECTS). Bone. 2022;154:116236. PubMed 34688942.
LeBoff MS, Greenspan SL, Insogna KL, et al. The clinician's guide to prevention and treatment of osteoporosis. Osteoporosis International. 2022;33(10):2049-2102. Bone Health & Osteoporosis Foundation. PubMed 35478046; open-access full text (PubMed Central).
US Preventive Services Task Force; Nicholson WK, Silverstein M, Wong JB, et al. Screening for Osteoporosis to Prevent Fractures: US Preventive Services Task Force Recommendation Statement. JAMA. 2025;333(6):498-508. PubMed 39808425; USPSTF recommendation page, final recommendation published January 14, 2025.
Wilding JPH, Batterham RL, Calanna S, et al. Once-Weekly Semaglutide in Adults with Overweight or Obesity (STEP 1). New England Journal of Medicine. 2021;384(11):989-1002. PubMed 33567185. Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide Once Weekly for the Treatment of Obesity (SURMOUNT-1). New England Journal of Medicine. 2022;387(3):205-216. PubMed 35658024.
ClinicalTrials.gov registry record, NCT07422987 — GLP-1RA Effects on Lean Mass and Bone Health in Midlife Women. University of Kansas Medical Center; prospective observational cohort, 50 women aged 35 to 60; estimated primary completion June 30, 2027.

This article is for information only and is not medical advice. Prescription weight-loss medication requires evaluation by a licensed clinician. See our medical disclaimer.