Meniscus Surgery vs Physical Therapy: What the Trials Actually Found

The short answer: for degenerative meniscal tears — the gradual, wear-related kind that shows up on MRI scans of middle-aged knees — high-quality randomised trials have repeatedly found that arthroscopic partial meniscectomy works no better than supervised exercise therapy. In one 146-patient trial, it worked no better than sham surgery. That is one of the most striking findings in modern orthopaedics. It is also narrower than it sounds: it does not apply to acute traumatic tears, bucket-handle tears, or a genuinely locked knee, where surgery may well be the right call.

First, the distinction that changes everything

Almost every argument about this evidence comes from people talking past each other because they mean different things by “meniscus tear.” There are broadly two situations, and they are not the same clinical problem.

Degenerative tears (what the trials studied)

These develop slowly, usually after 35, often without any single injury. The meniscus tissue frays and splits as part of the same wear process that produces knee osteoarthritis. Symptoms build over weeks or months. Crucially, degenerative tears are extremely common in people with no knee pain at all, which is why an MRI finding on its own does not establish that the tear is causing the symptoms.

Acute traumatic tears (what the trials did not study)

These follow a specific event — a twist under load, a tackle, a bad landing — typically in younger, athletic knees with otherwise healthy cartilage. Bucket-handle tears, where a torn fragment displaces into the joint, sit here too. This is a different injury in a different population, and the trials below deliberately excluded it.

Get assessed promptly, not eventually, if you have any of the following:

• A knee that is truly locked — objectively unable to fully straighten, not just stiff or clicky.
• Inability to bear weight, or a knee that gives way repeatedly.
• Significant trauma, immediate swelling within a few hours, or a pop at the moment of injury.
• Fever, redness, or heat in the joint, which needs urgent assessment for other reasons entirely.

Everything that follows is about the degenerative picture, and none of it replaces an assessment by an orthopaedic surgeon or physiotherapist who can actually examine your knee.

FIDELITY: the trial that compared surgery to fake surgery

The centrepiece is the Finnish Degenerative Meniscal Lesion Study, published in the New England Journal of Medicine in 2013. It is a multicentre, randomised, double-blind, sham-controlled trial in 146 patients aged 35 to 65 with symptoms of a degenerative medial meniscus tear and no knee osteoarthritis.

Every participant went to theatre and had a diagnostic arthroscopy. Only then were they randomised: half had the torn meniscus trimmed, half had the surgeon simulate the procedure — the same operating theatre, the same time under anaesthesia, the same sounds and sensations — and then close up without removing tissue. Patients and outcome assessors did not know which they had received.

At 12 months, in the intention-to-treat analysis, there were no significant between-group differences on any primary outcome:

• Lysholm score (0–100): improved 21.7 points after meniscectomy vs 23.3 after sham. Between-group difference −1.6 points (95% CI −7.2 to 4.0).
• WOMET score (0–100): improved 24.6 vs 27.1 points. Difference −2.5 points (95% CI −9.2 to 4.1).
• Knee pain after exercise (0–10): improved 3.1 vs 3.3 points. Difference −0.1 (95% CI −0.9 to 0.7).

Note the direction of every point estimate: they all slightly favour the sham group, and every confidence interval comfortably excludes a clinically meaningful advantage for surgery. Both groups improved a lot. Neither improved more than the other.

The follow-ups made the picture stronger, not weaker

At two years (Annals of the Rheumatic Diseases, 2018), the results held: WOMET difference −4.3 (95% CI −11.3 to 2.6), Lysholm −3.2 (95% CI −8.9 to 2.4). The investigators also ran the two subgroup analyses that surgeons most often invoke — patients with mechanical symptoms, and patients with unstable tears — and found no significant benefit in either.

At five years (British Journal of Sports Medicine, 2020), still no patient-relevant benefit: WOMET −1.7 (95% CI −7.7 to 4.3), Lysholm −2.1 (95% CI −6.8 to 2.6). There was a consistent signal in the other direction — a slightly greater risk of radiographic osteoarthritis progression in the surgery group (adjusted absolute risk difference of 13% for a Kellgren-Lawrence increase, 95% CI −2% to 28%), and more mechanical symptoms reported after meniscectomy (18%, 95% CI 5% to 31%).

METEOR: surgery plus rehab vs rehab alone

Published in the same journal in 2013, METEOR was a multicentre randomised trial of 351 patients aged 45 and over who had a meniscal tear plus mild-to-moderate osteoarthritis on imaging — a slightly different, more arthritic population than FIDELITY. They were assigned to arthroscopic partial meniscectomy with post-operative physical therapy, or to a standardised physical-therapy programme alone.

At six months, the mean improvement in WOMAC physical function was 20.9 points (95% CI 17.9 to 23.9) with surgery and 18.5 points (95% CI 15.6 to 21.5) with physical therapy — a mean difference of 2.4 points (95% CI −1.8 to 6.5), not statistically significant. Results at 12 months were similar, and adverse events did not differ significantly.

METEOR carries an important honest caveat that its own authors put in the conclusion: 30% of patients assigned to physical therapy alone crossed over and had surgery within six months. The trial was pragmatic by design — crossover was allowed at the patient’s and surgeon’s discretion. That tells you something real about the world: a meaningful minority of people on a conservative pathway will still end up wanting the operation.

Kise: exercise therapy head-to-head, with muscle data

The Norwegian trial published in The BMJ in 2016 randomised 140 adults (mean age 49.5) with MRI-verified degenerative medial meniscal tears to either 12 weeks of supervised exercise therapy alone or arthroscopic partial meniscectomy alone. Ninety-six percent had no definitive radiographic evidence of osteoarthritis.

At two years, the between-group difference in KOOS4 was 0.9 points (95% CI −4.3 to 6.1, P=0.72) — a difference the authors described as minute, with uncertainty small enough to exclude a clinically relevant advantage either way. At three months, thigh muscle strength had improved in the exercise group (P≤0.004) but not the surgical one. Nineteen percent of the exercise group crossed over to surgery over two years, with no additional benefit observed.

This is the practical asymmetry worth holding onto: exercise therapy bought something extra — quadriceps strength — that surgery did not.

What the systematic reviewers and guideline panels concluded

The 2022 Cochrane review pooled 16 randomised trials and 2,105 participants, of which four trials (380 participants) compared arthroscopic surgery to placebo surgery. Its verdict on the placebo comparison: high-certainty evidence of little or no difference in pain or function at three months. The pooled pain difference was 4.6 points better on a 0–100 scale (95% CI 0.02 to 9.0; 4 trials, 309 participants), and function differed by 0.1 points (95% CI −3.2 to 3.4). Whether arthroscopy slightly increases later knee replacement or osteotomy remained unresolved (RR 2.63, 95% CI 0.94 to 7.34).

The BMJ Rapid Recommendations panel — 18 members including orthopaedic surgeons, physiotherapists, a rheumatologist, methodologists and four patient representatives — issued a strong recommendation against arthroscopy in nearly all patients with degenerative knee disease, stating that further research is unlikely to change it. The panel found that under 15% of participants got a small or very small improvement in pain or function at three months, and that this benefit was not sustained at one year.

Read the guideline’s own scope carefully, because it draws exactly the line this article opened with. Its definition of degenerative knee disease explicitly excludes patients whose symptoms began recently after major knee trauma, and it names one possible exception to its own recommendation: people who are objectively unable to fully extend the knee — a true locked knee.

Limitations: where this evidence stops

Strong evidence is not unlimited evidence. Honest reading requires naming the edges.

1. Narrow populations. FIDELITY enrolled ages 35–65 with medial tears and no osteoarthritis; METEOR enrolled 45+ with existing osteoarthritis; Kise enrolled a mean age of 49.5. None speaks to a 22-year-old footballer with an acute lateral tear.
2. Crossover dilutes the contrast. Intention-to-treat analysis is the right analysis, but when 30% of METEOR’s conservative arm and 19% of Kise’s had surgery anyway, the comparison being measured is closer to “surgery now vs surgery if needed” than “surgery vs never.”
3. Follow-up is medium, not lifelong. The longest sham-controlled data run to five years. Effects on knee replacement decades later are not settled — the Cochrane estimate spans everything from no effect to a substantial increase.
4. Outcomes are patient-reported scores. Lysholm, WOMET, WOMAC and KOOS are validated, but they compress a complicated experience into a number, and small differences are hard to interpret.
5. The comparator was often good rehab. Supervised, progressive, 12-week exercise therapy is not the same as being handed a photocopied sheet of exercises. “No worse than surgery” assumes the rehab is actually delivered.
6. These trials tested meniscectomy, not meniscal repair. Repairing a repairable tear in a younger knee is a different operation with a different evidence base.

Funding and conflicts of interest

Worth knowing, because it is the first thing critics reach for. FIDELITY was funded by the Sigrid Juselius Foundation and other non-commercial sources (ClinicalTrials.gov NCT00549172); its authors declared no support from any organisation with an interest in the work. METEOR was funded by the US National Institute of Arthritis and Musculoskeletal and Skin Diseases (NCT00597012). The Kise trial’s authors declared no company support and no relevant financial relationships. The BMJ Rapid Recommendations panel included practising orthopaedic surgeons, and the journal judged that no panel member had a financial conflict of interest. In short, this is not industry-funded work, and surgeons were in the room.

How to use this in an actual conversation

The purpose of knowing this evidence is not to refuse an operation. It is to have a better-informed discussion with the surgeon and physiotherapist who are treating you. Reasonable questions include:

• Is my tear degenerative or traumatic, and what on my history and examination tells you that?
• Do I have a true mechanical block, or symptoms that merely feel mechanical?
• If I do a proper supervised 12-week rehab programme first, what would we expect, and what would make us change course?
• Given my imaging and age, which trial population do I most resemble?

For most middle-aged knees with a degenerative tear, a structured strength and mobility programme is the sensible first move — and unlike surgery, it also builds the quadriceps strength that protects the joint long term. If you want a starting point, our guide to science-backed exercises for knee pain at home covers the loading principles these trial protocols were built on.

The bottom line

For a degenerative meniscal tear in a middle-aged knee, the best available evidence — including a sham-controlled trial, a large pragmatic trial, a head-to-head exercise trial, a Cochrane review and a major guideline — says arthroscopic partial meniscectomy does not beat good conservative care. It does not say that no knee ever needs surgery. If you have had significant trauma, cannot bear weight, or your knee is genuinely locked, get assessed promptly; that scenario was explicitly outside these trials. For everything else, the trials give you permission to try rehab properly first, and to expect it to work about as well.

References

• Sihvonen R, et al. Arthroscopic partial meniscectomy versus sham surgery for a degenerative meniscal tear. N Engl J Med. 2013;369(26):2515-24. PubMed
• Sihvonen R, et al. Arthroscopic partial meniscectomy versus placebo surgery for a degenerative meniscus tear: a 2-year follow-up. Ann Rheum Dis. 2018;77(2):188-195. PubMed
• Sihvonen R, et al. Arthroscopic partial meniscectomy for a degenerative meniscus tear: a 5 year follow-up of the placebo-surgery controlled FIDELITY trial. Br J Sports Med. 2020;54(22):1332-1339. PubMed
• Katz JN, et al. Surgery versus physical therapy for a meniscal tear and osteoarthritis (METEOR). N Engl J Med. 2013;368(18):1675-84. PubMed
• Kise NJ, et al. Exercise therapy versus arthroscopic partial meniscectomy for degenerative meniscal tear in middle aged patients: randomised controlled trial with two year follow-up. BMJ. 2016;354:i3740. PubMed
• O’Connor D, et al. Arthroscopic surgery for degenerative knee disease (osteoarthritis including degenerative meniscal tears). Cochrane Database Syst Rev. 2022;3:CD014328. PubMed
• Siemieniuk RAC, et al. Arthroscopic surgery for degenerative knee arthritis and meniscal tears: a clinical practice guideline. BMJ. 2017;357:j1982. PubMed

This article is general information, not medical advice. Decisions about meniscal surgery belong to you, your orthopaedic surgeon and your physiotherapist, based on your own examination and imaging.