Knee osteoarthritis: why exercise is the treatment, not the thing to avoid

Osteoarthritis is still widely described as wear and tear, which is a poor metaphor and leads people to a poor conclusion. If a joint were a bearing wearing out, using it less would preserve it. Joints are not bearings.

Cartilage has no blood supply. It is nourished by the movement of synovial fluid, which is driven by compression and release as the joint loads and unloads. Immobilise a joint and cartilage thins. Astronaut and bed-rest studies show this clearly. Loading is not the enemy of cartilage; it is how cartilage is fed.

What we now understand osteoarthritis to be is a disease of the whole joint — cartilage, subchondral bone, synovium, ligaments and the surrounding muscle — with a low-grade inflammatory component. Not a passive erosion.

Pain and damage are only loosely related

This is the fact that most changes how people manage the condition, and it is well established.

Imaging studies repeatedly find substantial radiographic osteoarthritis in people with no pain, and severe pain in people with mild radiographic change. Roughly speaking, the correlation between X-ray severity and symptoms is weak. Cartilage itself has no nerve endings; the pain arises from synovium, bone, ligament and the surrounding tissues, and is modulated centrally.

The practical consequence is that a scan report saying "severe degenerative change" describes the picture and predicts your future rather poorly. People are frequently frightened out of activity by a radiology report, and the fear does more functional harm than the finding.

Why strength matters specifically

Quadriceps weakness is both a consequence of knee osteoarthritis and, on longitudinal evidence, a predictor of developing it. Muscle absorbs load. A weak quadriceps transmits more force through the joint at heel strike and provides less control through the movement.

Strengthening reduces pain and improves function in a large body of trials. It also has a plausible mechanism, which is unusual in this field.

The programme does not need to be complicated:

• Sit-to-stands from a chair, progressing to a lower seat, then to one leg. The most useful single exercise.
• Step-ups onto a low step, controlled on the way down — eccentric control is often the weakest element.
• Wall sits, isometric holds, which are useful when movement is provocative because they load the muscle without much joint travel.
• Hip abduction and extension. Hip weakness alters knee mechanics and is frequently the overlooked contributor.
• Calf raises, for the same reason at the other end.

Two or three sessions a week, progressively loaded. Twelve weeks before judging it.

The pain rule

People need permission to feel some discomfort, and a boundary for it. The convention used in physiotherapy is a numerical scale: pain up to about 5 out of 10 during exercise is acceptable, provided it settles within 24 hours and is not progressively worsening week to week.

Some soreness after starting is expected and is not damage. Sharp pain, giving way, locking, or a joint that swells and stays swollen are different, and warrant assessment.

Weight, said carefully

Load through the knee is roughly three to six times bodyweight during walking, so even modest weight loss changes joint loading meaningfully. Trials of weight loss in overweight people with knee osteoarthritis show consistent improvement in pain and function, and the combination of diet plus exercise outperforms either alone — the IDEA trial being the clearest example.

The mechanical explanation is incomplete, though. Osteoarthritis is also more common in the hands of people with obesity, where load is not the issue, which points toward metabolic and inflammatory contributions from adipose tissue.

What else has evidence, and what does not

Topical NSAIDs are effective for knee osteoarthritis and carry far less systemic risk than oral ones. They are under-used and are a reasonable first pharmacological step.

Paracetamol has a smaller effect than long assumed, and several guidelines have downgraded it.

Oral NSAIDs work, and carry gastrointestinal, renal and cardiovascular risk that matters more with each decade. Lowest dose, shortest duration, with gastric protection where indicated.

Corticosteroid injection gives short-term relief, typically a few weeks. Repeated injections have been associated with cartilage volume loss in at least one trial, so they are best used to open a window for rehabilitation rather than as an ongoing strategy.

Glucosamine and chondroitin have been extensively studied. The largest and best-conducted trials, including GAIT, found no benefit over placebo in the overall population. They are safe and mostly a transfer of money.

Arthroscopic surgery for degenerative knee disease — meniscal debridement, lavage — is recommended against by strong evidence, including sham-controlled trials. This one is worth knowing, because it is still offered.

Knee replacement is a genuinely effective operation for severe, refractory disease with poor quality of life. It is not an early option, and outcomes are better in people who enter surgery stronger — which is the argument for prehabilitation.

The reframe worth making

Osteoarthritis is a long-term condition to be managed rather than a countdown to a joint replacement. The people who do best are, consistently, those who kept moving — who adjusted the activity rather than abandoning it, who built the leg strength, and who did not accept the idea that a painful knee is a knee that should be rested.