A headline says a weight-loss medicine substantially reduced knee pain in a major trial. For someone living with osteoarthritis, the next question is natural: "Does that mean my cartilage is growing back?" It is a fair hope — and it runs ahead of what the studies actually measured. The best evidence says something genuinely good about pain and function, and something much quieter about the joint itself.
STEP 9: Strong Evidence for Pain Relief
The pivotal study is STEP 9, a 68-week, double-blind, randomized, placebo-controlled trial run at 61 sites in 11 countries, enrolling 407 adults with obesity and a clinical and radiologic diagnosis of moderate knee osteoarthritis with at least moderate pain. [1]
By week 68, average body weight had fallen 13.7% with semaglutide versus 3.2% with placebo, and the WOMAC knee-pain score — a 0-to-100 scale where higher means worse — improved by 41.7 points versus 27.5 points, both differences highly statistically significant. [1]
Note that those are points on the scale, not percentages — a distinction that matters when headlines round them into bigger-sounding numbers.
Physical function told the same story: SF-36 physical-function scores improved by 12.0 points with semaglutide versus 6.5 with placebo. [1]
On safety, serious adverse events occurred at similar rates in both groups, while stopping treatment because of side effects was more common with semaglutide (6.7% versus 3.0%), most often for gastrointestinal problems. [1]
The trial was funded by the drug's manufacturer — worth knowing — though its randomized, double-blind, placebo-controlled, multi-site design is a strong approach for estimating treatment effects. [1]
A 2026 Real-World Cohort Adds Nuance
A 2026 retrospective cohort followed 93 adults started on semaglutide in routine rheumatology care, 88 of whom completed a six-month follow-up, with no untreated comparison group. [2]
On average, patients lost 10.88 kg, pain on a visual scale fell by 2.47 points, total WOMAC scores fell by 22.50 points, and inflammatory markers and HbA1c also declined — all statistically significant within-patient changes. [2]
Weight loss remained independently associated with WOMAC improvement, and exploratory clustering hinted at distinct response patterns among patients. [2]
But the authors themselves are unusually explicit: without a concurrent control group, the design precludes causal attribution, and the observed patterns do not establish weight-independent pharmacologic effects or validated patient subtypes. [2]
In plain terms, patients got better while on the drug — but the study cannot isolate how much change arose from treatment, weight change, other care or the course of symptoms.
| STEP 9 outcome at 68 weeks | Semaglutide group | Placebo group |
|---|---|---|
| Average body-weight change | -13.7% | -3.2% |
| WOMAC pain change (0-100 scale) | -41.7 points | -27.5 points |
| SF-36 physical-function change | +12.0 points | +6.5 points |
| Serious adverse events | Similar rate to placebo | Similar rate to semaglutide |
| Stopped treatment for side effects | 6.7%, mostly gastrointestinal | 3.0% |
Pain Relief Is Not Cartilage Repair
Here is the distinction the headline misses: the reported outcomes concern symptoms, function and related measurements. Neither abstract reports a demonstration that knee cartilage regrew. [1][2]
The 2026 cohort explored weight change and inflammatory measurements, but its uncontrolled design could not establish the mechanism of improvement. Symptom relief alone cannot demonstrate structural regrowth. [2]
The prescribing information for semaglutide (Wegovy) lists indications for cardiovascular risk reduction, long-term weight reduction, and metabolic dysfunction-associated steatohepatitis — knee cartilage repair appears nowhere among them. [3]
Absence from the label is not proof of no structural benefit; it simply means this labeling does not support making that claim. By the same logic, these data do not show that semaglutide prevents knee replacement or changes anyone's need for surgery — surgical decisions belong with orthopedic clinicians, not pain-score headlines.
The Quality Gap Outside the Approved Supply
Separately, the FDA has warned that unapproved versions of GLP-1 drugs, including semaglutide, do not undergo agency review for safety, effectiveness, or quality before they are marketed. [4]
Compounded versions are not FDA-approved; the agency has received reports of dosing errors with compounded injectable semaglutide — some requiring hospitalization — as well as fraudulent products bearing false pharmacy labels and shipments that arrived warm or under-refrigerated. [4]
The FDA's position is that compounded drugs should be used only when an FDA-approved drug cannot meet the patient's medical need. [4]
For a condition like knee osteoarthritis, where the trial evidence concerns a defined formulation and supervised study, that quality gap is not a footnote — it is part of the risk calculation.
What This Means for Research Materials
These findings draw a clean line between three categories people often blur: the approved medicine tested in STEP 9, compounded or unapproved versions the FDA warns about, and laboratory research materials. A laboratory reagent can match its certificate — verified identity and high purity — and still lack the evidence and manufacturing controls needed for human use. Reconstitution for an assay cannot supply those missing assurances.
This does not presume that all reagents are contaminated or poorly tested. It is a claim that each category answers a different question, and research materials answer laboratory questions only — they should never be consumed or self-administered. Honesty about uncertainty also runs both ways: unproven cartilage effects are not disproven effects, and future imaging-based trials may tell us more. Until then, "less pain" is the finding; "repaired cartilage" is not.
Source note: Study findings are drawn from the original 2024 trial abstract and the 2026 cohort abstract. Current prescribing information and FDA safety information provide separate context. No cartilage-imaging benefit, surgical outcome or Chudik experiment is claimed; this is editorial source review, not medical review.