Research Square · Nov 4, 2026

Falls, fractures, lean mass and age with GLP-1receptor agonists and tirzepatide: meta-analysis of 162 placebo-controlled trials andage-subgroup data for adults aged 75 years orolder

By Ohe K

Background. Weight loss with glucagon-like peptide-1 receptor agonists (GLP-1 RAs) and tirzepatide includes loss of lean mass, raising concern about falls and fractures in older adults. Whether harm increases with participant age, and whether benefit persists at 75 years or older, has not been examined systematically. Methods. We extracted serious adverse events coded as "Fall", any fracture, and hip/femur fracture from all placebo-controlled, parallel-group randomised trials of GLP-1 RAs or tirzepatide in adults (treatment ≥ 12 weeks) with results posted on ClinicalTrials.gov (searched 31 October 2026), and searched PubMed (1 November 2026) for placebo-controlled trials reporting lean mass by dual-energy X-ray absorptiometry (DXA). Risk ratios (RR) and mean differences (MD) were pooled by random-effects models (REML, Hartung-Knapp), and effect modification by trial mean age was tested by meta-regression. Post hoc, we extracted results by participant age from FDA review documents and published subgroup analyses, and pooled placebo-controlled trials reporting 6-minute walk distance (6MWD) and bone-turnover markers. Results. In 162 trials (83,571 participants on GLP-1 RA or tirzepatide, 62,962 on placebo; trial mean ages up to 74 years), serious falls (k = 30; RR 0.91, 95% CI 0.77–1.07), any serious fracture (k = 76; RR 0.90, 0.80–1.01) and hip/femur fracture (k = 22; RR 1.04, 0.80–1.36) were not increased, and meta-regression showed no association with trial mean age (falls: RR per 10 years 0.92, 0.55–1.54; p = 0.75). Lean mass fell by 2.76 kg more than with placebo (9 trials, mean age 29–53 years; −3.78 to − 1.74), and by 4.12 kg in the four semaglutide or tirzepatide trials. Within SELECT (semaglutide 2.4 mg, n = 17,604), the fracture RR rose across age bands from 0.47 (0.25–0.91) at < 55 years to 1.30 (0.75–2.26) at ≥ 75 years (ratio per band 1.44, 1.11–1.87; p = 0.006); at ≥ 75 years, hip and pelvis fractures occurred in 17/703 vs 4/663 participants (RR 4.01, 1.36–11.85). At about 75 years or older (subgroups ≥ 70 to ≥ 75 years) in type 2 diabetes, HbA1c was lowered by 0.6–0.7 percentage points and the pooled hazard ratio for major adverse cardiovascular events was 0.80 (0.61–1.06); any serious adverse event was not more frequent with treatment in SELECT, LEADER or STEP-HFpEF. 6MWD improved by 15.2 m (8.6–21.9; k = 6). No placebo-controlled trial reported grip strength, gait speed or the Short Physical Performance Battery. Conclusions. Across trials, GLP-1 RAs and tirzepatide reduced lean mass without increasing serious falls or fractures, with no evidence that trial mean age modified this. Within the largest trial, however, the fracture RR increased with participant age, and hip and pelvis fractures were more frequent with semaglutide at 75 years or older. In adults with type 2 diabetes aged about 75 years or older, HbA1c is lowered and cardiovascular estimates point in the same direction as in younger adults without reaching significance. Trials in this age group that measure efficacy together with fat mass, lean mass, muscle strength, physical function and fractures are needed.

doi.org/10.21203/rs.3.rs-11601462/v1Open accessInterventions