A worn joint is the norm of mature age

Degenerative changes in the knee after thirty-five or forty are not a pathology but an expected age-related finding, like grey hair or wrinkles.

A meta-analysis of 63 studies, almost five thousand people, published in the British Journal of Sports Medicine in 2018, showed that in people with no complaints and no injuries such changes are widespread. Cartilage defects are found in 11% of people under 40 and already in 43% after 40, meniscal tears in 4% under 40 and in 19% after 40.¹

A study of 230 knees in ordinary sedentary adults with a median age of 44 (3-tesla MRI, Skeletal Radiology, 2020) took the point to its limit: some abnormality in the knee joints was found in 97% of those examined, meniscal tears in 30%.²

And finally the Framingham study (New England Journal of Medicine, 2008), which every orthopedist cites. Among people aged 50 to 90 meniscal tears were very common, but 61% of those with a torn meniscus had no pain, no aching and no stiffness.³

The missing link between the imaging picture and the clinical state

A systematic review by Bedson and Croft (BMC Musculoskeletal Disorders, 2008) brought together dozens of studies. Among people with knee pain, radiographic osteoarthritis is found in 15 to 76% of cases, while among those whose osteoarthritis is visible on imaging, pain is present in only 15 to 81%.⁴ The spread is enormous, the correlation unconvincing.

Both when radiographic changes are present and when they are absent, a large role in the pain is played by so-called central sensitization, a re-tuning of the nervous system’s own pain-processing machinery. A study in Arthritis & Rheumatism (2013) showed that signs of central sensitization are especially pronounced in patients with a high level of pain in the absence of moderate or severe changes on X-ray.⁵

Why do knees hurt?

Like any pain syndrome, knee pain has many causes at once. But two of them I would call the main ones.

The first is the state of the muscles. Weakness of the quadriceps is one of the earliest signs, appearing even before the pain. A review in Frontiers in Physiology (2021) states plainly that muscle weakness is characteristic of patients with knee osteoarthritis and predicts loss of function better than pain or joint space narrowing.⁶

The second cause is chronic subclinical inflammation. Modern rheumatology has effectively rewritten the definition of osteoarthritis. A review in Nature Reviews Rheumatology (2016) puts it this way: osteoarthritis was long considered a degenerative disease of cartilage, but accumulated data show that inflammation plays a key role, and this inflammation is chronic, comparatively low-grade and driven by innate immunity.⁷ The same belongs here: age-related smouldering inflammation, inflammaging, and inflammation of the synovial membrane. Synovitis is linked both to more pronounced pain and to faster cartilage loss.⁸

Treating knee pain

Here the evidence is almost unequivocal, and it clearly runs against the usual surgical logic.

The Norwegian OMEX study (BMJ, 2016) enrolled 140 middle-aged patients with a degenerative meniscal tear, and in 96% there were no clear radiographic signs of osteoarthritis.⁹

Ten-year follow-up found no difference either. The ESCAPE study (five-year follow-up) confirmed it: arthroscopic meniscal resection has no clinically meaningful advantage over exercise therapy.¹⁰

The strictest form of evidence is comparison with a sham operation. The Finnish FIDELITY study (New England Journal of Medicine, 2013) compared a real meniscal resection with an imitation operation, in which incisions are made but the meniscus is not touched, in 146 people aged 35 to 65. The result was the same. The ten-year follow-up of the same study, published in April 2026, went further: surgery gave no advantage over sham surgery, and in the surgery group there was a trend toward worse long-term outcomes, including more symptoms, more pronounced progression of osteoarthritis and more frequent repeat procedures.¹¹

This is not news but a confirmation of earlier studies. Moseley’s work (New England Journal of Medicine, 2002) already showed that in knee osteoarthritis the outcomes after arthroscopic lavage or debridement are no better than after a placebo procedure.¹² The Kirkley study (New England Journal of Medicine, 2008) repeated this in another group: adding arthroscopy to physical and drug therapy gave no additional benefit.¹³

This is not theory. The GLA:D program, grown from these principles, gathered data on 28,370 patients in Denmark, Canada and Australia: people with symptomatic osteoarthritis of the knee and hip show significant improvements in pain intensity, function and quality of life after a course of intensive physiotherapy.¹⁵ An important detail: in this program the diagnosis is made clinically, without imaging, and only about 10% of patients with knee osteoarthritis proceed to joint replacement within two years.¹⁶

The choice: should you have an MRI for knee pain?

If MRI changes are present in half of healthy peers, and the best results come from strengthening muscles and treating inflammation rather than from surgery, the question arises: what role does imaging itself play. The answer is an uncomfortable one: most often an excessive role.

A register-based study from Norway gives figures every patient should know.

Meanwhile the number of scans keeps growing: in Australia, between 2003 and 2017, the number of knee MRIs increased sevenfold.¹⁷

For a typical patient the diagnosis is made without any equipment at all. In people over 45 with activity-related pain and morning stiffness under half an hour, imaging is usually not needed unless there are atypical symptoms.¹⁸

When a scan is justified, the X-ray comes first, not the MRI, because a plain radiograph is considered the first-line examination in osteoarthritis.¹⁹

The harm here is not abstract. An MRI will almost certainly find “degeneration” or a “tear” that is present in many symptom-free people of the same age. The patient takes fright, the doctor gains a reason to intervene, then comes arthroscopy, which, as we have seen, works no better than placebo and over the long term may worsen the outcome. A scan does not treat the knee. It sets off a chain of anxiety and unnecessary interventions.