Dr Matthew BroadheadOrthopaedic Surgeon · Sydney & Coffs Harbour

Journal Club · July 2026 · updated 11 November 2026

How long does a knee replacement last?

In the Australian registry's 2026 report, about 92% of total knee replacements done with implants still in use had not been revised at 20 years. Age at surgery makes a large difference.

How long will it last? It is one of the most common questions before knee replacement surgery, and a fair one. Two sources answer it: a 2019 review in The Lancet that pooled the long-term evidence from around the world, and the 2026 report of the Australian joint replacement registry, which has the most recent figures and covers the implants used today.

What the studies looked at

A knee replacement is counted as having lasted for as long as it has not needed a further operation to replace part or all of it. That second operation is called a revision.

The Lancet review (2019). Researchers in Bristol pooled every report they could find of knee replacements followed for more than 15 years, relying most on national registries, which record nearly every operation done in a country. The registry data covered roughly 300,000 total knee replacements, chiefly from Australia and Finland.

The Australian registry (2026). The Australian Orthopaedic Association National Joint Replacement Registry has recorded more than 98% of joint replacements done in Australia since 1999. Its 2026 report covers operations up to the end of 2025, and for total knee replacement done for osteoarthritis it reports on 775,654 operations. It limits most of its analyses to implants that were still being used in 2025, which it calls modern prostheses. This year it also estimated, for the first time, the chance of needing a revision over the rest of a person’s life.

Both looked separately at total knee replacement, which resurfaces the whole joint, and partial knee replacement, which resurfaces one side of it.

What they found

Total knee replacement in Australia, with implants in current use:

  • At 10 years, about 5 in 100 had been revised (4.6%).
  • At 15 years, about 6 in 100 (6.2%).
  • At 20 years, 8 in 100 (8.0%). Put the other way, about 92% had not needed a revision 20 years on.

Age at the time of surgery made a large difference. At 20 years:

  • under 55 at surgery: 16 in 100 had been revised
  • 55 to 64: about 10 in 100
  • 65 to 74: about 6 in 100
  • 75 and over: about 4 in 100

Partial knee replacement was revised more often: about 11 in 100 at 10 years and about 26 in 100 at 20 years.

The Lancet review, with older implants, estimated that about 82% of total knee replacements and about 70% of partial knee replacements were still in place at 25 years. The two sources tell a consistent story for the total knee.

Over a lifetime. The registry’s new estimate is that the younger a person is at the time of surgery, the greater the chance of needing a revision at some point, simply because there are more years ahead. Its modelled lifetime risk after a total knee replacement was about 31% for people aged 46 to 50 at surgery, about 12% for people aged 66 to 70, and about 2% for people 86 and over. It was higher in men than in women, and higher in people who are obese.

What it means for you

For most people, a total knee replacement done today can be expected to last a long time. On the Australian figures, about 9 in 10 are still in place 20 years later.

Your age matters more for a knee than for a hip. For someone having a knee replaced in their seventies, it is likely to be the only operation that knee needs. For someone in their early fifties, the chance of a second operation at some point is real, around one in six within 20 years on these figures, and that is part of an honest conversation about timing. It is rarely a good reason to put up with years of severe pain and disability.

A partial knee replacement is a smaller operation with a quicker recovery, and it is more likely to need revising later, often because arthritis develops in the rest of the knee. The registry’s own summary is that it can still be the right choice for some people. That is a decision to make with your surgeon, knowing both sides of the trade.

When a total knee replacement does need to be redone, the most common reason now is infection, followed by loosening.

None of this tells you what will happen to your own knee. That depends on your age, your weight, your general health and the implant used.

A caveat worth knowing

The registry’s figures are for implants still in use in 2025. Designs that were withdrawn along the way, often because they did not perform well, are left out, and the registry says openly that this makes the results look better.

The registry does not yet report a figure at 25 years, and its lifetime estimates are projections from a model, less certain the further ahead they look. They are not predictions for any one person.

Both sources count only revision. A knee that is still in place but painful counts as a success, and so does one that would have been revised had the person been well enough for surgery. As always, general findings like these are a starting point for a discussion, not a substitute for advice about your own knee.

Assessing this studyA standardised appraisal for clinicians and registrars: study design, strength and validity.
Study design
Two sources read together. Evans 2019 is a systematic review and meta-analysis of case series and national registry reports with more than 15 years of follow-up; the endpoint is all-cause revision. The AOANJRR 2026 Annual Report is a national registry cohort recording over 98% of joint replacements in Australia from September 1999 to December 2025, reporting cumulative percent revision by Kaplan-Meier estimate, and, new this year, a modelled lifetime risk of revision.
Level of evidence
OCEBM Level 2 (prognosis) for both: a systematic review of observational registry data, and a very large inception cohort. Neither is randomised. The lifetime risk figures are modelled projections and sit below the observed 20-year figures in certainty.
Are the results valid?
Evans 2019: a pre-specified protocol and broad search, weighted towards national registries, but the 25-year estimates rest on few registries (chiefly Australia and Finland), on older devices, and are unadjusted for age. AOANJRR 2026: near-complete national capture, linked to death records, with narrow confidence intervals. Its main limitation is deliberate: nearly all analyses are restricted to "modern prostheses", meaning those still in use in 2025. Designs that were withdrawn, often because they performed poorly, are left out, so the figures describe current implants and are better than the experience of everyone who had a knee replaced in that period. The Kaplan-Meier estimates are unadjusted and weighted towards older patients.
What are the results?
AOANJRR 2026, primary total knee replacement for osteoarthritis, modern prostheses, 775,654 procedures (Table KT12): cumulative percent revision 4.6% (95% CI 4.5–4.6) at 10 years, 6.2% (6.1–6.3) at 15 years and 8.0% (7.8–8.2) at 20 years. By age at operation, 20-year revision was 16.0% under 55, 10.2% at 55 to 64, 6.3% at 65 to 74 and 3.8% at 75 and over (Table KT15). Unicompartmental replacement, 54,120 procedures (Table KP18): 11.1% at 10 years and 26.3% (25.4–27.2) at 20 years. Modelled lifetime risk of revision after total knee replacement fell from 30.6% for patients aged 46 to 50 to 12.5% at 66 to 70 and 2.0% at 86 and over (Table N39). Evans 2019: pooled 25-year survival 82.3% (81.3–83.2) for total and 69.8% (67.6–72.1) for unicompartmental replacement.
Do they apply to our patients?
The AOANJRR figures are Australian, current, and directly applicable to counselling here. They describe how implants in use today have performed so far; no 25-year figure is presented in the 2026 report. Age at operation changes the estimate fourfold between the youngest and oldest groups, so the overall figure should not be quoted to a younger patient without the age table. Revision is a narrow definition of failure; a painful knee that has not been revised counts as surviving.
Bottom line
With implants in current use, about 92% of total knee replacements done for osteoarthritis in Australia had not been revised at 20 years (AOANJRR 2026), consistent with the 82% at 25 years in the 2019 Lancet review of older implants. The figure is strongly age-dependent (16% revised at 20 years under 55, under 4% at 75 and over) and should be quoted with the patient's age in mind.

Framework: study design graded with the Oxford Centre for Evidence-Based Medicine (OCEBM) Levels of Evidence; validity and applicability appraised in the CASP tradition.

This is a plain-language summary of published research, provided for general education. It is not medical advice and does not describe Dr Broadhead's own results. Whether any finding applies to you depends on your individual circumstances, so please discuss your care with your treating team.

← All Journal Club entries

CallBookEmail