Short answerIn one trial, knee replacement followed by non-surgical care gave more pain relief and better function at one year than non-surgical care alone, with more serious complications.
Probably helps a little
What it is
Total knee replacement removes the worn surfaces of the knee and replaces them with an implant. It is offered when arthritis is moderate to severe and non-surgical care has not been enough. Non-surgical care in the trial meant exercise, education, diet advice, shoe insoles and pain medicine.
What the research shows
enrolled 100 people with moderate to severe knee arthritis who were suitable for a knee replacement. Half had a knee replacement and then 12 weeks of non-surgical care. The other half had the 12 weeks of non-surgical care only. At 12 months, the surgery group improved by about 32 points on a 100 point knee score, compared with about 16 points in the non-surgical group.
Both groups improved. Only 13 of the 50 people in the non-surgical group had a knee replacement within the year.
showed the same pattern, and about one in three people in the non-surgical group had a knee replacement by then. This is one small trial. People knew which treatment they had, and there was no sham surgery group, so the gap may be overstated.
What the Australian joint registry shows
In the Australian joint registry, about 95 in 100 total knee replacements done for arthritis had not needed another operation at 10 years, and about 92 in 100 at 20 years . Knees in younger people were redone more often, about 16 in 100 by 20 years in people under 55, compared with about 4 in 100 in people aged 75 and over. Infection is now the most common reason for another operation, followed by loosening and instability.
Six months after surgery, about 86 in 100 people in the registry were satisfied or very satisfied with their knee, and about 4 in 100 felt it was worse.
Risks and effort
Serious complications were more common with surgery. The trial recorded 24 serious adverse events in the surgery group and 6 in the non-surgical group. They included blood clots in the leg veins and stiffness that needed a manipulation under anaesthetic. Recovery takes weeks to months of rehabilitation.
What usually works instead
A supervised program of exercise, weight loss where it applies, and pain relief gave a large gain on its own. See Does exercise work for knee arthritis?, Does losing weight help knee arthritis? and Osteoarthritis of the Knee. See also Knee replacement.
Questions people ask
Do I need to try non-surgical care first?
The trial authors advised a full, supervised non-surgical program before surgery. Many people in that group did well and did not need an operation.
Will a knee replacement last?
The trial followed people for one year, so it cannot say. In the Australian joint registry, about 92 in 100 knee replacements for arthritis had not needed another operation after 20 years. The registry estimates the chance of needing another operation at some point in life is about 16 in 100 for people having surgery at 61 to 65, about 12 in 100 at 66 to 70, about 9 in 100 at 71 to 75 and about 7 in 100 at 76 to 80 .
Who gains most?
The trial did not answer this. People with severe pain and loss of function that has not eased with non-surgical care are the usual candidates.
Can I manage without surgery if my knee arthritis is severe?
Many people do. In the trial, everyone had moderate to severe arthritis and was suitable for a replacement. Those who started with 12 weeks of supervised non-surgical care also improved, and 13 of the 50 had a replacement within a year. Surgery gave more improvement on average, so the choice depends on how much pain and loss of function you are living with.
What are the downsides of a knee replacement?
In the trial, serious complications were more common in the surgery group, with 24 events compared with 6 in the non-surgical group. They included blood clots in the leg veins and stiffness that needed treatment under anaesthetic. Recovery takes weeks to months of rehabilitation.
Sources
Select a source to read what the study or guideline found, in plain words, with a link to the original.
Last reviewed 24 October 2026. Next review due October 2027.
