Reviewed for medical accuracy by Dr Evelyne Jonniaux, Medical Advisor
Last updated 17 September 2026
“A joint replacement lasts 10 to 15 years”: that figure has no source
We found no paper, no registry and no guideline that establishes it. What is actually measured says something else. The registry review by Evans (The Lancet, 2019) covers 215,676 hip replacements and 299,291 knee replacements: at 25 years, 57.9% of hips and 82.3% of knees are still in place. And in contemporary practice, a hip implanted in 2014 has a 10-year revision probability of 2.7%, meaning more than 97% still in place (National Joint Registry, 2025 report). These figures describe populations, not one joint. Noria Health arranges a full orthopaedic work-up near Brussels within 24 to 48 hours.
How long they actually last
| Implant, registry data | 15 years | 20 years | 25 years |
|---|---|---|---|
| Total hip | 89.4% | 70.2% | 57.9% |
| Total knee | 93.0% | 90.1% | 82.3% |
| Unicompartmental knee | 76.5% | 71.6% | 69.8% |
These come from Evans (The Lancet, 2019), drawn from national registries, which the authors judge less prone to bias than series published by the operating teams themselves. Case series give markedly more flattering numbers, 77.6% at 25 years for the hip, and it is precisely that gap which argues for using the registries.
A 2026 update (Pentland, The Lancet) restricted to modern bearing surfaces, across 1.9 million hips, observes 93.6% survival at 20 years. The 25 and 30-year values in that work are model extrapolations, not observations, and we present them as such.
Revision rates, year by year
| National Joint Registry, 2025 report | 1 year | 10 years | 15 years | 20 years |
|---|---|---|---|---|
| Hip, all cases (1,682,998 implants) | 0.79% | 3.63% | 5.81% | 8.38% |
| Hip, excluding metal-on-metal | 0.78% | 2.90% | 4.76% | 7.14% |
| Knee, all types (1,805,000 implants) | 0.49% | 3.82% | 5.49% | 7.58% |
| Cemented knee, the commonest construct | 0.39% | 2.69% | 3.74% | 5.28% |
| Uncemented metal-on-metal hip | 1.08% | 17.76% | 22.51% | 25.95% |
That last row is why the type of implant is a question worth asking. Metal-on-metal bearings, widely implanted in the 2000s, carry a 10-year revision rate six times the average. They are no longer used, but people are still living with them.
Hips and knees do not give the same result
This is the mismatch least well explained to patients. In one Ontario cohort, same method, same team (Bourne, Clin Orthop Relat Res, 2010), overall satisfaction is 89% after hip replacement and 81% after knee replacement. Would have the operation again: 96% against 89%. Expectations met: 78% against 70%.
The English national programme of patient-reported outcome measures, 2023-24 data, finds the same gap at scale: 84.8% of hip patients report themselves much better, against 73.5% of knee patients.
And persistent pain after knee replacement is real. Beswick (BMJ Open, 2012) put it around 20% in the better-quality studies. The same team’s update gives 12.6% at 12 months and 14.6% at 24 months. Citing the 2012 figure without the update would be an error by omission.
The strongest predictor of dissatisfaction after knee replacement is not the technique: it is unmet expectations, with a risk multiplied by 10.7. What was said before the operation weighs more than almost anything else.
What a replacement adds, compared with not operating
One randomised trial has compared knee replacement against well-conducted non-surgical treatment. Skou (NEJM, 2015) allocated 100 patients eligible for a replacement between surgery followed by 12 weeks of non-surgical treatment, and 12 weeks of non-surgical treatment alone, meaning supervised exercise, education, dietary advice, insoles and analgesics.
| What is measured | The result |
|---|---|
| KOOS4 improvement at 12 months, replacement | 32.5 points |
| KOOS4 improvement at 12 months, non-surgical | 16.0 points |
| Adjusted mean difference | 15.8 points (95% CI 10.0 to 21.5) |
| Serious adverse events, replacement | 24 |
| Serious adverse events, non-surgical | 6, p = 0.005 |
| Non-surgical arm eventually operated at 1 year | 26% |
| Non-surgical arm eventually operated at 2 years | 32% |
The replacement does twice as well, and that is a real result. But two patients in three who were eligible for a replacement had not had one two years later after well-run non-surgical treatment, and the gain is paid for in serious adverse events, four times as many.
In real-world populations the average effect is more modest. Ferket (BMJ, 2017), across 4,498 participants followed for nine years, measures an improvement attributable to knee replacement of just 1.70 points on the physical component of the SF-12, with an uncertainty interval from 0.26 to 3.57. The key point of that work is that the benefit grows sharply as baseline function is worse. Operating on a mildly disabling knee gains little; operating on a severely disabling one gains a great deal.
The degenerative meniscus, and the trial nobody likes to quote
Arthroscopy for a degenerative meniscal tear remains common. The FIDELITY trial (Sihvonen, NEJM, 2013) compared partial meniscectomy against sham surgery, double-blind, in 146 patients aged 35 to 65 without osteoarthritis.
| Improvement at 12 months | Meniscectomy | Sham surgery | Difference |
|---|---|---|---|
| Lysholm score | 21.7 | 23.3 | -1.6 (-7.2 to 4.0) |
| WOMET score | 24.6 | 27.1 | -2.5 (-9.2 to 4.1) |
| Pain after exercise | 3.1 | 3.3 | -0.1 (-0.9 to 0.7) |
Both groups improve a great deal, by 22 to 27 points. It is the size of that improvement in the sham group, not any absence of improvement, that is the finding. At five years (Br J Sports Med, 2020) there is still no difference on patient-reported outcomes, and one counter-intuitive result: more mechanical symptoms in the operated group, absolute risk difference 18% with an interval from 5% to 31%.
The argument that the knee locks was tested separately within the same trial (Ann Intern Med, 2016). Among patients with catching or locking before surgery, the risk difference for locking at follow-up is 0.07 with an interval from -0.08 to 0.22. Locking does not identify the patients who benefit from the procedure.
What the studies do not allow us to claim
That one must operate early or the functional loss becomes irrecoverable is not demonstrated. The largest work on the question, 89,996 patients, does document measurable deterioration during the wait, but could not observe any relationship with the result after the operation.
That some given percentage of patients gain no benefit from arthroplasty has no single source: five ways of measuring give five different figures. We do not adopt a single number.
Finally, the American guideline often invoked to fix the timing of surgery explicitly excludes surgery from its scope. The attribution is wrong.
In practice, if this concerns you
Three questions structure the discussion. How disabling is it today, since the benefit of the operation grows with that. Is it a hip or a knee, since the satisfaction rates differ and it is better to know beforehand. And which implant and bearing surface are proposed, since the spread between constructs runs from 2.2% to 17.8% revision at 10 years.
Noria Health arranges a full orthopaedic work-up near Brussels within 24 to 48 hours: clinical and functional examination, weight-bearing radiographs, assessment of general condition and operative risk factors, and a surgical opinion where the examination justifies it. The report is issued within 24 hours, follow-up is structured at D14, D30 and D90, and surgery is considered only around day 10 if it is needed.
What Noria Health does not do: propose a replacement on a radiograph alone. A painful joint with fever, or sudden loss of use after an impact, are matters for the emergency services in your country, immediately, not for an appointment.
Further reading
- What waiting costs in orthopaedics
- Spine: at 50, 80% of people with no pain have a degenerate disc
- Pain that waits
Sources
- Evans JT, Evans JP, Walker RW, Blom AW, Whitehouse MR, Sayers A. How long does a hip replacement last? The Lancet, 2019;393:647-654. https://pubmed.ncbi.nlm.nih.gov/30782340/
- Evans JT, Walker RW, Evans JP, Blom AW, Sayers A, Whitehouse MR. How long does a knee replacement last? The Lancet, 2019;393:655-663. https://pubmed.ncbi.nlm.nih.gov/30782341/
- Pentland V, et al. Survival of modern total hip replacements. The Lancet, 2026;407:855-866. https://pubmed.ncbi.nlm.nih.gov/41763743/
- National Joint Registry. 22nd Annual Report, 2025, chapter 3, tables 3.H5 and 3.K5. https://www.njrcentre.org.uk/
- Bourne RB, Chesworth BM, Davis AM, Mahomed NN, Charron KDJ. Patient satisfaction after total knee arthroplasty, who is satisfied and who is not? Clin Orthop Relat Res, 2010;468:57-63. https://pubmed.ncbi.nlm.nih.gov/19844772/
- Bourne RB, Chesworth B, Davis A, Mahomed N, Charron K. Comparing patient outcomes after THA and TKA, is there a difference? Clin Orthop Relat Res, 2010;468:542-546. https://pubmed.ncbi.nlm.nih.gov/19760472/
- Anakwe RE, Jenkins PJ, Moran M. Predicting dissatisfaction after total hip arthroplasty. J Arthroplasty, 2011;26:209-213. https://pubmed.ncbi.nlm.nih.gov/20462736/
- NHS England. Patient Reported Outcome Measures, final 2023-24 data. https://digital.nhs.uk/data-and-information/publications/statistical/patient-reported-outcome-measures-proms/final-2023-24-data
- Beswick AD, Wylde V, Gooberman-Hill R, Blom A, Dieppe P. What proportion of patients report long-term pain after total hip or knee replacement? BMJ Open, 2012;2:e000435. https://pubmed.ncbi.nlm.nih.gov/22357571/
- Skou ST, Roos EM, Laursen MB, et al. A randomized, controlled trial of total knee replacement. N Engl J Med, 2015;373:1597-1606. https://pubmed.ncbi.nlm.nih.gov/26488691/
- Skou ST, Roos EM, Laursen MB, et al. Total knee replacement and non-surgical treatment of knee osteoarthritis, 2-year outcome. Osteoarthritis Cartilage, 2018;26:1170-1180. https://pubmed.ncbi.nlm.nih.gov/29723634/
- Ferket BS, Feldman Z, Zhou J, Oei EH, Bierma-Zeinstra SMA, Mazumdar M. Impact of total knee replacement practice. BMJ, 2017;356:j1131. https://pubmed.ncbi.nlm.nih.gov/28336564/
- Sihvonen R, Paavola M, Malmivaara A, et al. Arthroscopic partial meniscectomy versus sham surgery for a degenerative meniscal tear. N Engl J Med, 2013;369:2515-2524. https://pubmed.ncbi.nlm.nih.gov/24369076/
- Sihvonen R, Paavola M, Malmivaara A, et al. Arthroscopic partial meniscectomy for a degenerative meniscus tear, a 5 year follow-up. Br J Sports Med, 2020;54:1332-1339. https://pubmed.ncbi.nlm.nih.gov/32855201/
- Sihvonen R, Englund M, Turkiewicz A, Jarvinen TLN. Mechanical symptoms and arthroscopic partial meniscectomy. Ann Intern Med, 2016;164:449-455. https://pubmed.ncbi.nlm.nih.gov/26856661/
See also: all our articles on musculoskeletal and pain.
Have your case reviewed. Noria Health arranges a specialist consultation near Brussels and confirms feasibility within 24 to 48 hours. Open a case.