GLP-1 drugs and ankle OA: new cohort data

A retrospective cohort study published in Foot & Ankle Specialist on 17 August 2026 asked a question most GLP-1 research has not: do these drugs change what happens to ankle joints?^[1] The answer, for obese patients, was yes. For patients with type 2 diabetes, the picture was more complicated.

What the study did

Le and colleagues at Johns Hopkins drew from the TriNetX Analytics Network, a real-world database covering millions of patient records from 2016 through 2020. They identified two populations: adults with obesity (BMI 30 kg/m2 or above) and adults with type 2 diabetes. Within each group they compared patients who had received at least two prescriptions for a GLP-1 receptor agonist, separated by six months or more, against matched controls who had not. GLP-1 agents included in the definition were semaglutide, liraglutide, dulaglutide, exenatide, and lixisenatide. After propensity-score matching the final cohorts were 3,089 obese patients per group and 8,117 diabetic patients per group.^[1]

Outcomes tracked from six months after the index date were: ankle joint injections, ankle arthroscopy, and total ankle arthroplasty (TAA). The study did not track ankle arthrodesis (fusion), so the surgical spectrum it covers is injections, scope procedures, and replacement. For context on how semaglutide is regulated and who can prescribe it, see the peptide page.

What it found in the obese cohort

In patients with obesity, sustained GLP-1 use was associated with a meaningfully lower likelihood of two out of three ankle interventions.^[1] The hazard ratio for ankle joint injection was 0.8 (95% CI 0.6 to 0.9), roughly a 20 percent lower rate versus non-users. The hazard ratio for total ankle arthroplasty was 0.5 (95% CI 0.3 to 0.9), roughly a halving in the rate of ankle replacement. Arthroscopy showed no statistically significant difference between groups.

The metabolic backdrop: GLP-1 users in the obese cohort achieved a mean BMI reduction of 2.7 kg/m2 over the follow-up period, versus 1.6 kg/m2 in controls (p less than 0.01). Whether the lower ankle intervention rate reflects weight offloading on the joint, anti-inflammatory effects of GLP-1 receptor activation, or both is not resolved by this study design.

What it found in the diabetic cohort

The diabetic cohort produced a different and counterintuitive result.^[1] GLP-1 use was associated with a higher rate of ankle arthroscopy: hazard ratio 1.9 (95% CI 1.2 to 3.2). Joint injection and total ankle arthroplasty showed no statistically significant association in either direction.

BMI reductions were also observed in diabetic GLP-1 users (mean drop of 2.1 kg/m2 versus 1.7 kg/m2 in controls), but HbA1c improvement was slightly better in the non-user group (a drop of 6.0 percentage points versus 5.5 in GLP-1 users; p less than 0.01). The elevated arthroscopy rate in the diabetic cohort may reflect better overall healthcare engagement among patients on a complex pharmacological regimen. The authors are careful not to claim a causal explanation from retrospective data. For background on semaglutide in metabolic disease, the peptide page covers the full trial programme.

How this sits alongside the knee OA evidence

The ankle findings arrive after a landmark in the knee OA literature. The STEP 9 randomised trial, published in the New England Journal of Medicine in October 2024, enrolled 407 adults with obesity and symptomatic knee osteoarthritis. Semaglutide 2.4 mg weekly produced a mean body-weight reduction of 13.7 percent versus 3.2 percent on placebo at 68 weeks.^[2] The KOOS pain subscale score improved by 41.8 points on semaglutide versus 27.5 points on placebo, a separation of roughly 14 points on a scale where 12 points is considered clinically important.^[2]

The Le et al. study is a real-world complement to that trial evidence: not a controlled efficacy test, but a look at whether the pattern holds at the level of clinical decision-making in a large database. The knee data showed a pain benefit; the ankle data suggests a reduction in the procedures that typically follow failed conservative management. Both observations involve the same class of agents, anchored by semaglutide as the most-studied member.

The surgical risk angle

A separate line of evidence has been building around GLP-1 use before joint surgery. A 2024 analysis in the Journal of Arthroplasty examined patients who took semaglutide before total hip arthroplasty and found lower rates of postoperative prosthetic joint infection and 90-day readmission versus controls.^[3] The mechanism proposed is primarily metabolic: reduced adipose tissue, lower systemic inflammation, and better glycaemic control before a major procedure. The Le et al. ankle study does not directly test perioperative use, but the population overlap is real: patients at highest ankle OA burden are often the same patients being evaluated for surgical fitness.

Why ankle OA is a different problem from knee OA

Ankle OA is less common than knee OA, and its aetiology is different. Most ankle OA cases are post-traumatic (a fracture or ligament injury years earlier), whereas knee OA is largely primary and load-related. That matters for extrapolating the semaglutide knee data: the pain relief in STEP 9 was partly driven by weight reduction, and the ankle joint is not subjected to the same peak loading forces per kilogram of body weight as the knee. Le et al. did not stratify by post-traumatic versus primary ankle OA, so whether GLP-1 effects on ankle joints are mostly about weight or partly about direct joint biology remains open.

A 2026 narrative review in Rheumatology International examined direct GLP-1 receptor signalling in joint tissue and found preclinical evidence that GLP-1 receptors are expressed in chondrocytes and synovial fibroblasts, with receptor activation reducing pro-inflammatory cytokine production.^[4] Whether that mechanism is operative at the ankle at therapeutic doses is not yet established in human data. For the broader weight-independent mechanism debate, see the semaglutide osteoarthritis explainer on this site.

What we do not yet know

Le et al. has a standard set of retrospective limitations: TriNetX captures prescriptions, not adherence; the follow-up window is short relative to ankle OA's natural history; and the analysis cannot rule out confounding by indication. The divergent finding in the diabetic cohort (more arthroscopy, not less) is harder to explain and warrants prospective investigation. The study does not include ankle arthrodesis, the most common surgical endpoint in severe ankle OA. Ankle OA is not currently listed in any regulatory indication for semaglutide, liraglutide, or any other GLP-1 agent.

What this means if you have ankle osteoarthritis

This article is for educational and journalistic purposes. GLP-1 receptor agonists are prescription medicines, authorised in most countries for type 2 diabetes and weight management in adults, not for osteoarthritis. The ankle OA evidence base is at an early stage: one retrospective cohort, a related but not identical knee trial, and mechanistic data from animal models. If you or a patient has ankle osteoarthritis alongside obesity or type 2 diabetes, the appropriate route is a conversation with an orthopaedic specialist or rheumatologist. Country-specific prescribing status for semaglutide is on the semaglutide regulation pages on this site.