GLP-1s and Knee Osteoarthritis

This is the comorbidity that most reliably changes how someone’s day feels, and it is the one people are least likely to raise, because knee pain gets filed under ageing rather than under something treatable.

The mechanism is load

The knee is a hinge carrying your body weight through a small contact area. Biomechanical work has repeatedly shown that compressive force through the knee is a multiple of body weight with each step — the exact multiplier varies with gait and activity, and the direction is not in dispute. Walking, stairs and rising from a chair all amplify it further.

So a reduction in body weight produces a proportionally larger reduction in the force driving pain, inflammation and cartilage wear. That is why weight loss works here in a way it does not work for every condition it gets recommended for: the mechanism is mechanical and direct.

There is a metabolic component too — osteoarthritis has an inflammatory dimension, and adipose tissue is metabolically active — but load is the larger and more certain part.

What the trial showed

STEP 9, published in the NEJM in 2024, randomised adults with obesity and knee osteoarthritis to semaglutide or placebo. Over 68 weeks it found significantly greater reduction in pain scores alongside the weight loss, with improvement in physical function.

That sits on top of older lifestyle evidence. The IDEA trial found that diet plus exercise produced greater pain reduction and lower knee compressive force than either alone, at losses of around 10%.

Two things worth noting. The pain improvement in these trials is meaningful but not total — this is a condition being made more manageable, not cured. And diet plus exercise beat diet alone, which is the finding most relevant to what you do next.

What weight loss does not do

Cartilage does not grow back. Osteoarthritis involves structural change to cartilage, subchondral bone and the joint capsule, and nothing available reverses it.

This matters practically, because the pain improvement can be substantial enough to feel like repair. People stop the physiotherapy, resume running, and are surprised when the knee objects. The joint is less loaded, not restored. Protect it accordingly.

The muscle problem, which is the whole complication

Here is where this intersects with everything else on this site.

Quadriceps strength is among the better predictors of function in knee osteoarthritis. Weak quadriceps means less shock absorption, worse stability and more pain, independent of weight.

And a substantial share of the weight lost on these drugs is lean tissue. So it is entirely possible to arrive lighter, with a knee carrying less load and supported by less muscle — and to feel not much better. That is the failure mode, and it is avoidable.

Protein at 1.2–1.6 g/kg and resistance training twice a week is the answer, as it is in muscle is the whole game. What changes here is exercise selection.

Training a knee that hurts

Osteoarthritis is not a reason to avoid resistance training. It is a reason to choose the load carefully, and this is worth a session with a physiotherapist rather than guesswork.

The general framing in cardio, steps and heart rate still applies — the aerobic work matters — and here the mode selection matters as much as the amount. Pain during exercise that settles within 24 hours is generally acceptable, and 2 sessions a week is the target; pain that persists or swells the joint means the load was wrong.

When the joint pain is not osteoarthritis

A joint that becomes red, hot and exquisitely tender over hours — classically the big toe, often overnight — is a different problem, and rapid weight loss can provoke it. That is GLP-1s and gout, and the distinction matters because the treatment is completely different.

If surgery is on the table

Worth raising explicitly with your surgeon rather than assuming.

Many centres consider weight in decisions about elective joint replacement, and substantial loss can change eligibility, anaesthetic risk and how rehabilitation goes. Some people find the pain improves enough that surgery is deferred; others still need it and arrive in better condition.

Two practical notes if you do proceed. Tell the anaesthetist you take a GLP-1 — a stomach that empties slowly is a specific consideration, covered in GLP-1s, surgery and procedures. And the muscle you keep is the muscle you rehabilitate with, which makes the training argument above a pre-operative one rather than a general-wellbeing one.

What I would ask for

• A physiotherapy referral at the start, not after the knee complains
• A protein target written down, and resistance training twice a week from week one
• A conversation about rate of loss, since faster is not better for muscle
• If surgery is a live question, a surgical opinion that accounts for where your weight is heading, rather than one based on today’s number