Searches for "Ozempic bone loss" roughly quadrupled over the past year. The record behind them is thinner and more specific than the headlines: one GLP-1 label counts fractures, a handful of small trials measured bone density, and the large 2026 studies all turn on one variable — whether the person has type 2 diabetes. This page sets those out with the absolute numbers, every one read from the label or paper on 2 October 2026.
For lean mass (a separate question that is often blurred with this one) see GLP-1 muscle loss; for the full side-effect record see GLP-1 side effects.
The one label that counts fractures
Four current labels were read for this page. Only Wegovy's (semaglutide for weight management, label effective 2026-06-18) reports fractures. The Ozempic (2026-06-01), Mounjaro (2026-08-27) and Zepbound (2026-08-28) labels contain no fracture or bone-density wording at all.
The Wegovy numbers come from SELECT, the cardiovascular outcomes trial in adults with heart disease and overweight or obesity, and from the MASH liver trial:
| Group (label wording) | On Wegovy | On placebo | Extra per 100 people |
|---|---|---|---|
| Women, SELECT — hip and pelvis fractures | 1% (24/2,448) | 0.2% (5/2,424) | ≈ 0.8 |
| Aged 75 and over, SELECT — hip and pelvis fractures | 2.4% (17/703) | 0.6% (4/663) | ≈ 1.8 |
| Adults in the MASH trial — all fractures | 4.4% (2.6 per 100 patient-years) | 3.3% (2.0 per 100 patient-years) | ≈ 1.1 |
Extra per 100: glp1ledger arithmetic from the label's counts (24/2,448 = 0.98%, 5/2,424 = 0.21%; 17/703 = 2.42%, 4/663 = 0.60%).
Two ways of reading the same row give very different impressions. Relatively, women in SELECT had almost five times the rate of hip and pelvis fractures on the drug. Absolutely, that is about one extra fracture for every 130 women treated over the trial's follow-up — and about one in 55 among people aged 75 and over. Hip fractures in older people carry real consequences, so neither framing should be dropped.
What the label does not say: it reports these as adverse reactions, not as a tested hypothesis. SELECT was not designed to measure bone, did not scan bone density, and the label gives no figure for men or for the trial as a whole. The geriatric section adds only that people 75 and over "reported more hip and pelvis fractures" on Wegovy than on placebo.
The trials that measured bone density
Bone mineral density (BMD) is measured with a DXA scan. Very few GLP-1 trials have done that with bone as a stated outcome.
| Study | Who | What was compared | Bone result |
|---|---|---|---|
| Jensen 2024, JAMA Netw Open (randomised, 52 weeks) | 195 adults with obesity, no diabetes, after an 8-week low-calorie diet | Liraglutide 3.0 mg, exercise, both, or placebo | Liraglutide alone: hip BMD −0.013 g/cm² and spine −0.016 g/cm² against exercise alone, at similar weight loss. Liraglutide + exercise: no change against placebo |
| Dinkla 2025, Front Aging (pilot, 20 weeks) | 20 adults averaging 73 years, with prediabetes or diabetes | Semaglutide 1 mg + counselling vs counselling | No significant difference in whole-body BMD or bone-turnover markers; the authors note BMD was consistently lower and turnover higher in the semaglutide group |
| Liu 2026, J Clin Endocrinol Metab (retrospective, matched) | 255 semaglutide or tirzepatide users with repeat DXA scans (92% women, mean age 64) and 255 non-users | Users vs matched non-users, median 17 months | Both groups lost hip BMD. Without diabetes, users lost more: about −1% against −0.6% a year. With diabetes, no difference. More weight lost went with more hip bone lost |
The liraglutide figure is small in the units it is reported in. The trial's liraglutide group started at a mean total-hip BMD of 1.03 g/cm², so −0.013 g/cm² is roughly 1.3% of baseline (glp1ledger arithmetic) — over one year, and on a drug weaker than the current weekly ones. Nobody has published the same design with semaglutide 2.4 mg or tirzepatide.
The large studies, and the variable that flips them
Bigger numbers come from health-record studies, which cannot prove cause but can count fractures across hundreds of thousands of people. Three published in 2026 all used the TriNetX network, and the answer depends on who was studied and what the drug was compared with.
| Study | Population | Comparator | Result |
|---|---|---|---|
| Hamad 2026, JAMA Netw Open | 133,606 adults 50–90 with type 2 diabetes | DPP-4 inhibitors | Fewer fragility fractures on GLP-1 drugs: HR 0.79 (hip/femur 0.70, vertebral 0.68); one fewer fracture per 126 treated over 3 years |
| Hamad 2026, same paper, stratified | With vs without type 2 diabetes | DPP-4 inhibitors | With diabetes HR 0.91; without diabetes HR 1.13 — more fractures |
| Huang 2026, Diabetes Obes Metab | Obesity, matched pairs, with and without type 2 diabetes | Empagliflozin, glipizide or usual care | With diabetes: fewer major osteoporotic fractures (HR 0.69–0.84 by comparator). Without diabetes: no significant association |
Put next to the Weill Cornell scan data and the Wegovy label, the pattern is consistent: every signal of harm comes from people without type 2 diabetes, or from older women, and every "protective" result comes from people with diabetes compared against other diabetes drugs. Diabetes itself raises fracture risk, and better glucose control may help bone, so a benefit there need not carry over to weight-loss use. This is the same lesson as with cancer findings on this site: the comparator can flip the sign, so it has to be named.
The widest summary so far — a 2026 meta-analysis of 60 studies including 46 randomised trials and about 1.25 million people (Beaudart 2026, Drugs) — found no effect on bone density or fractures when it used each study's most-adjusted estimate, but a consistent fall in lean mass. Its authors rated the certainty of the evidence as low.
Tirzepatide: an absence, not an all-clear
Neither tirzepatide label mentions bone. Tirzepatide also acts on the GIP receptor, which sits on bone-forming and bone-resorbing cells, and reviewers have flagged bone density as a question for the newer multi-receptor drugs (see maridebart cafraglutide). The only tirzepatide bone data located for this page is the Weill Cornell study, which pooled it with semaglutide. A label that says nothing is a label that says nothing.
What is not established
- Whether the drug harms bone beyond what the same weight loss would. One small trial says liraglutide alone did; no trial has tested semaglutide 2.4 mg or tirzepatide the same way.
- Whether lower density turns into fractures in younger adults. The fracture signal in the Wegovy label is in women and people 75 and over.
- Whether exercise prevents it with the newer drugs. The Copenhagen result is for liraglutide.
- What happens to bone after stopping. No study located here followed bone after discontinuation.
People at higher fracture risk — older women, people with osteoporosis or past fractures — have the most to weigh here, and the decision about scanning, exercise or treatment belongs with their own clinician. Hormonal context matters as well; see GLP-1 drugs and menopause. Protein intake during weight loss is covered on protein and GLP-1, and the wider long-run record on long-term side effects.
This page reports what labels and studies state. It is not advice on whether to start, continue or stop any medicine.
Sources
- Wegovy prescribing information (Novo Nordisk), DailyMed set ID ee06186f-2aa3-4990-a760-757579d8f77b, effective 2026-06-18 — Adverse Reactions (Fractures) and Geriatric Use. dailymed.nlm.nih.gov, read 2 October 2026.
- Ozempic prescribing information, DailyMed set ID adec4fd2-6858-4c99-91d4-531f5f2a2d79 (2026-06-01); Mounjaro, d2d7da5d-ad07-4228-955f-cf7e355c8cc0 (2026-08-27); Zepbound, 487cd7e7-434c-4925-99fa-aa80b1cc776b (2026-08-28) — searched for fracture and bone-density wording, none found. Read 2 October 2026.
- Jensen SBK, et al. Bone health after exercise alone, GLP-1 receptor agonist treatment, or combination treatment. JAMA Netw Open 2024;7(6):e2416775. doi:10.1001/jamanetworkopen.2024.16775 (PMID 38916894; full text PMC11200146 for baseline BMD).
- Dinkla L, et al. Bone mineral density and turnover response to GLP-1 receptor agonists in older adults. Front Aging 2025;6:1691007. doi:10.3389/fragi.2025.1691007 (PMID 41393101).
- Liu Y, et al. Skeletal effect of semaglutide and tirzepatide in patients with increased risk of fractures. J Clin Endocrinol Metab 2026;111(7):1959–1966. doi:10.1210/clinem/dgag052 (PMID 41655226).
- Hamad CD, et al. Glucagon-like peptide-1 receptor agonists and fragility fracture risk in type 2 diabetes. JAMA Netw Open 2026;9(7):e2625141. doi:10.1001/jamanetworkopen.2026.25141 (PMID 42496972).
- Huang YN, et al. Associations of semaglutide with skeletal outcomes in people with obesity, with and without type 2 diabetes: a target trial emulation. Diabetes Obes Metab 2026;28(7):5834–5847. doi:10.1111/dom.70786 (PMID 42010367).
- Beaudart C, et al. GLP-1 receptor agonists and musculoskeletal outcomes: a systematic literature review and meta-analysis. Drugs 2026;86(10):1661–1691. doi:10.1007/s40265-026-02365-3 (PMID 42730869).