Skip to content

Condition guide · Knee and lower limb

Knee Osteoarthritis

What knee osteoarthritis is, what the X-ray does and does not tell you, and what the guidelines actually recommend — including several widely sold treatments the evidence does not support.

Evidence OARSI 2019 guideline· ACR/AF 2019 guideline· Cochrane review (139 trials, 12,468 participants)

In plain words. The ends of the bones inside your knee are capped with cartilage, a smooth layer that lets the joint glide. In osteoarthritis that cartilage thins, the bone underneath changes, and the lining of the joint becomes irritated. The knee gets stiff, aches, and objects to being used the way it used to be.

The single most useful thing to know is that how bad the knee looks on a scan and how much it hurts are two different questions. Plenty of people have impressive-looking scans and very little pain. Plenty of people have modest scans and a great deal of pain. Treatment is aimed at the second question.

Key points

  • The two major international guidelines agree on what comes first: education and a structured land-based exercise programme, with weight management where relevant. [1][2]
  • Scan findings are a poor guide to symptoms. In a general population sample, 61% of people found to have a meniscal tear on MRI had no pain, aching or stiffness in the previous month. [8]
  • Arthroscopy is strongly recommended against for nearly all people with degenerative knee disease — with or without a meniscal tear on imaging. [5]
  • Platelet-rich plasma injection performed no better than salt water in a blinded trial of 288 people, on either pain or cartilage volume. [6]
  • Exercise helps, but the most recent Cochrane review is candid that the size of the benefit is of uncertain clinical importance. We say so rather than round it up. [3]
  • Knee osteoarthritis is common in India: a meta-analysis of 16 community surveys estimated a prevalence of 20.24% (95% CI 11.90 to 30.13). [9]

What is actually happening

Osteoarthritis is often described as "wear and tear", as though the joint were a brake pad with a fixed mileage. That description does not survive contact with the data. It is better understood as a condition of the whole joint — cartilage, the bone beneath it, the lining, the ligaments and the muscles that control the knee — in which the joint's ability to repair and adapt is outpaced by the demands placed on it.

Two things point away from the simple mechanical story. First, structural damage and symptoms track each other poorly, which is hard to explain if pain were simply a readout of cartilage thickness. [8] Second, in an 18-month randomised trial, intensive weight loss reduced both the compressive force through the knee and a marker of systemic inflammation, and the group that lost weight and exercised had less pain than either alone. [4] Load and inflammation both appear to be part of it.

The practical consequence is encouraging: several of the things that drive symptoms — muscle strength, activity tolerance, body weight, how the knee is loaded — can be changed. The X-ray cannot.

What it feels like

  • Pain that comes on with activity and settles with rest, particularly stairs, squatting, kneeling and getting up from a low chair.
  • Stiffness after sitting still, which eases within a few minutes of moving.
  • A sense that the knee is unreliable — it may feel as if it could give way, usually because the muscles controlling it have weakened.
  • Grinding or crunching. This is extremely common and, on its own, is not a sign of damage.
  • Swelling that comes and goes, often after a day of more walking than usual.
  • Good weeks and bad weeks. Fluctuation is normal and does not mean the joint is deteriorating each time.

How it is diagnosed

Mainly by listening and examining. Age, the pattern of the pain, what makes it worse, how long the morning stiffness lasts, what the knee looks like and how it moves, and how strong the surrounding muscles are, together give the diagnosis in most cases.

What imaging can and cannot settle

Imaging is useful when something specific needs ruling in or out. It is much less useful as a way of explaining pain, because the findings are so common in people who have no symptoms at all.

In a Framingham population sample of 991 people aged 50 to 90, chosen without regard to whether their knees hurt, meniscal tears on MRI were found at almost the same rate in those with and without knee symptoms. [8]

Meniscal tear on MRI — how often it is found in people who have no knee symptoms

Meniscal tears on MRI are almost as common in people without knee symptoms as in people with them Bar chart of data from a Framingham population sample of 991 people aged 50 to 90. Among people with osteoarthritis visible on X-ray, a meniscal tear was found in 63 percent of those with knee pain, aching or stiffness on most days, and in 60 percent of those without those symptoms. Among people with no osteoarthritis on X-ray, a tear was found in 32 percent of those with symptoms and 23 percent of those without. Knee pain on most days No knee symptoms Osteoarthritis visible on X-ray 63% 60% No osteoarthritis on X-ray 32% 23% Percentage of people in whom a meniscal tear or meniscal destruction was seen on MRI. Framingham population sample, n=991, ages 50 to 90, selected without regard to knee problems. [8]

Across that whole sample, 61% of the people who had a meniscal tear reported no pain, aching or stiffness at all in the previous month. [8] The clinical practice guideline on knee arthroscopy puts the conclusion bluntly: meniscus tears are common, usually incidental findings, and unlikely to be the cause of knee pain, aching or stiffness. [5]

None of this means imaging is useless. It means a tear on a report is not, by itself, an explanation, and still less an indication for surgery.

Illustration of a knee joint affected by osteoarthritis, showing thinning of the cartilage between the femur and tibia.

Cartilage change is what the scan shows. It is a poor guide to how much pain someone has or how well they will respond, which is why management is built around what the joint can do rather than what it looks like.

How physiotherapy and rehabilitation help

Exercise does not regrow cartilage. What it does is improve the strength, control and tolerance of the joint and the muscles around it, so that the same knee does more with less pain. That is a real result and it is worth being precise about, because a treatment sold as repairing the joint is being oversold.

The 2023 EULAR recommendations for non-pharmacological core management of hip and knee osteoarthritis set out eight evidence-based recommendations, agreed at between 9.2 and 9.8 out of 10 by the task force. In order, they cover an individualised multicomponent management plan; information, education and self-management; exercise with adequate tailoring of dosage and progression; mode of exercise delivery; maintaining a healthy weight and weight loss; footwear, walking aids and assistive devices; work-related advice; and behaviour change techniques. [10]

Read that list as a description of what a good service actually looks like. Only one of the eight items is the exercise itself. The rest is the scaffolding that decides whether the exercise gets done — which is also what the guideline's research agenda identifies as the open problem: adherence, uptake and impact on work.

Which exercise, and how much it matters

A 2025 network meta-analysis in the BMJ pooled 217 randomised trials and 15,684 participants to compare exercise modalities directly. On moderate-certainty evidence, aerobic exercise probably produces large improvements in pain in the short term (SMD -1.10, 95% CI -1.68 to -0.52) and mid term (-1.19, -1.59 to -0.79), in function at mid term (1.78, 1.05 to 2.51), in gait performance at mid term (0.85, 0.55 to 1.14) and in quality of life at short term (1.53, 0.47 to 2.59). [11]

Other modalities also came out well on moderate certainty: mind-body exercise for short-term function (0.88, 0.03 to 1.73), neuromotor exercise for short-term gait performance (1.04, 0.51 to 1.57), and both strengthening (0.86, 0.53 to 1.18) and mixed exercise (1.07, 0.68 to 1.46) for mid-term function. [11]

The practical reading is that walking and other aerobic activity is not the poor relation of a gym-based strengthening programme — on this evidence it is at least as well supported, and it is far easier to sustain. Anyone told that only supervised strengthening will help is being given a narrower account than the evidence supports.

If your knee gave way at the time of an injury rather than becoming painful gradually, the relevant page is ACL injury rehabilitation — and note that an ACL tear raises the long-term risk of the osteoarthritis described here.

What treatment involves

The OARSI guideline designates two things as Core Treatments for knee osteoarthritis — the things everybody should be offered, on which anything else is built: arthritis education, and a structured land-based exercise programme, with or without dietary weight management. [1] The ACR and Arthritis Foundation guideline makes strong recommendations for the same territory: exercise; weight loss for people who are overweight or obese; self-efficacy and self-management programmes; tai chi; use of a cane; and tibiofemoral bracing. [2]

In practice a programme is built around strengthening the muscles that control the knee and hip, restoring the range the joint has lost, and rebuilding tolerance to the activities that matter to you — stairs, standing at work, walking to the market, sitting cross-legged. It is graded: the load starts at a level the knee accepts and goes up as it adapts. Most of the work is done by you, between appointments.

Two points that are worth saying out loud. Exercise for an arthritic knee is often uncomfortable at first, and mild discomfort that settles within a day is not damage. And the Cochrane review found no difference in effect between the different types of exercise studied, and no relationship between improvement and the total number of sessions prescribed. [3] The programme you will actually do beats the theoretically optimal one you will not.

What the evidence supports — and what it does not

Supported

  • Structured land-based exercise — a Core Treatment in OARSI and a strong recommendation in ACR/AF. [1][2] Compared with no treatment or usual care, pain improved by a mean of 13.14 points on a 0–100 scale (95% CI 10.36 to 15.91; 56 studies, 4,184 people) and physical function by 12.53 points (95% CI 9.74 to 15.31). [3]
  • Weight management combined with exercise — in an 18-month randomised trial of 454 adults, the diet-plus-exercise group lost a mean 10.6 kg (11.4% of body weight) and reported less pain (3.6 on a 0–20 scale) than diet alone (4.8) or exercise alone (4.7). [4]
  • Arthritis education and self-management — a Core Treatment in OARSI; self-efficacy and self-management programmes carry a strong recommendation in ACR/AF. [1][2]
  • Topical NSAIDs — strongly recommended for knee osteoarthritis in both guidelines (OARSI Level 1A). [1][2]
  • A cane, and tibiofemoral bracing where the inner or outer compartment is affected — both strongly recommended in ACR/AF. [2]

Not supported

  • Knee arthroscopy — a strong recommendation against, in nearly all people with degenerative knee disease, with or without imaging evidence of osteoarthritis, meniscal tears or mechanical symptoms. Fewer than 15% of trial participants got even a small or very small improvement at three months, and that was gone by one year. [5]
  • Platelet-rich plasma injection — against saline placebo over 12 months, the difference in pain was −0.4 points (95% CI −0.9 to 0.2; P=0.17) and in medial tibial cartilage volume −0.2% (95% CI −1.9% to 1.5%). 29 of 31 secondary outcomes showed no significant difference. [6]
  • Repeated corticosteroid injection as a long-term strategy — 40 mg triamcinolone every 12 weeks for two years produced greater cartilage loss than saline (−0.21 mm vs −0.10 mm; difference −0.11 mm, 95% CI −0.20 to −0.03) and no significant difference in pain. [7]
  • Oral and transdermal opioids — strongly not recommended (OARSI Level 5). [1]
  • Oral NSAIDs in people with cardiovascular disease or frailty — not recommended in that group, whatever the knee is doing. [1]

Two honest complications. A single intra-articular corticosteroid injection for short-term pain relief is a different proposition from repeating it for years: both guidelines still place intra-articular corticosteroid among their recommended options for knee osteoarthritis. [1][2] The trial above is evidence against the long-term repeated schedule, not against ever using one. And the two guidelines disagree about paracetamol: OARSI conditionally recommends against it, while ACR/AF conditionally recommends it. [1][2] When guidelines of that quality disagree, the honest answer is that the effect is small enough to be argued about.

Pain felt in the knee is not always coming from the knee. Hip osteoarthritis commonly refers pain to the thigh and knee, and shares most of its management — see hip osteoarthritis.

How certain is this?

Evidence grade: Moderate. The direction of travel is not in serious doubt — two independent guideline panels, using GRADE, put education, exercise and weight management first, and the negative findings on arthroscopy and platelet-rich plasma come from large, blinded, placebo-controlled work. [1][2][5][6]

What is less certain is the size of the benefit from exercise, and we would rather say so than sell it. The 2024 Cochrane review included 139 trials and 12,468 participants and rated the certainty of evidence as low to moderate. Most trials were at unclear or high risk of bias, and in 94% of them participants were not blinded and knew which treatment they were getting. When the authors compared their results against established thresholds for a minimal important difference, the confidence intervals either failed to reach those thresholds or spanned both a clinically important and a clinically unimportant improvement. Their own conclusion is that the benefits were "of uncertain clinical importance". [3]

How the exercise effect compares with the threshold for a difference patients notice

The confidence intervals for the effect of exercise span the minimal important difference Two confidence intervals plotted on a 0 to 20 point scale. Pain improved by 13.14 points, 95 percent confidence interval 10.36 to 15.91, against a minimal important difference of 12 points. Physical function improved by 12.53 points, 95 percent confidence interval 9.74 to 15.31, against a minimal important difference of 13 points. In both cases the interval crosses the threshold, so the result is compatible with both an important and an unimportant benefit. 0 5 10 15 20 Points of improvement on a 0–100 scale Pain 56 studies threshold 12 13.14 Physical function 54 studies threshold 13 12.53 Exercise versus no treatment, usual care or limited education, measured immediately after the programme. Bars show 95% confidence intervals. Both cross the threshold for a difference patients notice. [3]

Read that chart honestly and it says: exercise probably helps, the average effect sits close to the threshold at which people notice a difference, and the data cannot rule out that for some people the benefit is smaller than that. That is still the best-supported first-line treatment available, and it is one of the few with no meaningful downside.

What to expect

Knee osteoarthritis is a long-term condition. There is no treatment currently available that restores worn cartilage, and any claim to the contrary should be examined carefully. What can change, often substantially, is how much pain the knee produces and how much you can do with it.

Improvement from an exercise programme is usually measured in weeks to months, not days, and it is not linear — flare-ups happen and do not undo the progress. The gains also depend on continuing: the trials measured people who were still exercising. Where weight is a factor, the trial evidence points to a substantial reduction being what shifts symptoms, and to combining it with exercise rather than choosing between them. [4]

A minority of people with severe, persistent symptoms that have not responded to good non-surgical care go on to consider joint replacement. That is a conversation with an orthopaedic surgeon, and it is a genuinely different operation from arthroscopy, which is the procedure the guideline above recommends against. [5]

When knee pain is not osteoarthritis — seek urgent assessment

Seek immediate medical attention if you have:

  • A hot, red, very swollen knee, especially with fever or feeling unwell — a joint infection is a medical emergency.
  • A knee that suddenly locks and cannot be straightened fully, and stays that way.
  • Inability to put weight through the leg after an injury, or an obvious deformity.
  • Calf pain, swelling or warmth, particularly after surgery, immobility or a long journey.
  • Pain that is severe at night, wakes you consistently, or comes with unexplained weight loss or fever.
  • New numbness, pins and needles, or weakness in the leg or foot.

This list is not exhaustive. If you are worried about a symptom, seek professional advice.

Common questions

Will exercise wear my knee out faster?

No. That is the central worry patients bring, and it is the opposite of what the guidelines conclude. Structured land-based exercise is a Core Treatment in the OARSI guideline and a strong recommendation in ACR/AF; if it accelerated joint damage neither panel could have made that call. [1][2] Discomfort during and shortly after exercise that settles is expected, not a warning sign.

My X-ray says "severe". Does that mean I need surgery?

Not on its own. The severity on imaging and the severity of symptoms are only loosely related. [8] Decisions about surgery are made on how much the knee limits your life after a proper trial of non-surgical treatment, not on the grade written on a report.

What about PRP or stem cell injections?

For platelet-rich plasma there is now a well-conducted answer: in 288 people randomised to three weekly PRP injections or saline, there was no significant difference in pain or in cartilage volume at 12 months, and 29 of 31 secondary outcomes showed no difference either. [6] We have not cited a comparable trial of stem cell injection here, and would not recommend paying for a treatment on the strength of evidence we cannot show you.

Does glucosamine or chondroitin help?

Neither of the two guidelines cited here lists glucosamine among its recommended treatments for knee osteoarthritis, and the ACR/AF conditional recommendation for chondroitin sulfate is specific to hand osteoarthritis, not the knee. [1][2] They are not dangerous; the case that they work for the knee is not made in these guidelines.

Is a steroid injection a bad idea?

A single injection for a bad flare is a recognised option in both guidelines. [1][2] What the evidence argues against is repeating it on a schedule for years: two years of injections every 12 weeks produced more cartilage loss than saline and no significant pain benefit. [7]

How long before I notice a difference?

Weeks to months with a programme you actually do, rather than days. The trials that showed benefit measured people at the end of supervised programmes that typically ran for weeks, and the review found no relationship between improvement and how many sessions were prescribed. [3] Consistency matters more than intensity.

References

  1. Bannuru RR, Osani MC, Vaysbrot EE, et al. OARSI guidelines for the non-surgical management of knee, hip, and polyarticular osteoarthritis. Osteoarthritis and Cartilage. 2019 Nov;27(11):1578–1589. doi:10.1016/j.joca.2019.06.011 PMID 31278997 Clinical practice guideline
  2. Kolasinski SL, Neogi T, Hochberg MC, et al. 2019 American College of Rheumatology/Arthritis Foundation Guideline for the Management of Osteoarthritis of the Hand, Hip, and Knee. Arthritis Care & Research. 2020 Feb;72(2):149–162. doi:10.1002/acr.24131 PMID 31908149 Clinical practice guideline
  3. Lawford BJ, Hall M, Hinman RS, et al. Exercise for osteoarthritis of the knee. Cochrane Database of Systematic Reviews. 2024 Dec 3;12(12):CD004376. doi:10.1002/14651858.CD004376.pub4 PMID 39625083 Systematic review
  4. Messier SP, Mihalko SL, Legault C, et al. Effects of intensive diet and exercise on knee joint loads, inflammation, and clinical outcomes among overweight and obese adults with knee osteoarthritis: the IDEA randomized clinical trial. JAMA. 2013 Sep 25;310(12):1263–1273. doi:10.1001/jama.2013.277669 PMID 24065013 Randomised controlled trial
  5. Siemieniuk RAC, Harris IA, Agoritsas T, et al. Arthroscopic surgery for degenerative knee arthritis and meniscal tears: a clinical practice guideline. BMJ. 2017 May 10;357:j1982. doi:10.1136/bmj.j1982 PMID 28490431 Clinical practice guideline
  6. Bennell KL, Paterson KL, Metcalf BR, et al. Effect of intra-articular platelet-rich plasma vs placebo injection on pain and medial tibial cartilage volume in patients with knee osteoarthritis: the RESTORE randomized clinical trial. JAMA. 2021 Nov 23;326(20):2021–2030. doi:10.1001/jama.2021.19415 PMID 34812863 Randomised controlled trial
  7. McAlindon TE, LaValley MP, Harvey WF, et al. Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial. JAMA. 2017 May 16;317(19):1967–1975. doi:10.1001/jama.2017.5283 PMID 28510679 Randomised controlled trial
  8. Englund M, Guermazi A, Gale D, et al. Incidental meniscal findings on knee MRI in middle-aged and elderly persons. New England Journal of Medicine. 2008 Sep 11;359(11):1108–1115. doi:10.1056/NEJMoa0800777 PMID 18784100 Population-based cross-sectional study
  9. Hazra S, Chandra Kumar U, Mandal P, et al. Prevalence of knee osteoarthritis in India: a systematic review and meta-analysis of population-based studies. Indian Journal of Orthopaedics. 2025 Nov;59(11):1785–1796. doi:10.1007/s43465-025-01520-4 PMID 41245277 Systematic review and meta-analysis
  10. Moseng T, Vliet Vlieland TPM, Battista S, et al. EULAR recommendations for the non-pharmacological core management of hip and knee osteoarthritis: 2023 update. Annals of the Rheumatic Diseases. 2024 May 15;83(6):730–740. doi:10.1136/ard-2023-225041 PMID 38212040 Clinical practice guideline
  11. Yan L, Li D, Xing D, et al. Comparative efficacy and safety of exercise modalities in knee osteoarthritis: systematic review and network meta-analysis. BMJ. 2025 Oct 15;391:e085242. doi:10.1136/bmj-2025-085242 PMID 41093618 Systematic review and network meta-analysis

About this guide

Written by
Dr Satish Kumar (PT)MPT · Director & Head of Department · APARC Centre for Neurorehabilitation, Pitampura, Delhi
Reviewed by
Dr Dharam Pandey (PT)MPT; PhD · Chief Editor · Director & Head of Department · Department of Physiotherapy & Rehabilitation Science · not the author
Chief Editor
Dr Dharam Pandey (PT)MPT; PhD
Evidence grade
ModerateSee "How certain is this?"
Last reviewed
15 August 2026Next review due 15 August 2028

If you need assessment

This page explains. It does not diagnose.

Knee pain in an adult has several possible causes, and the exercise programme that helps one knee can aggravate another. Which muscles are weak, how the joint moves, and what you need the knee to do can only be established in person. Physiotherapist India publishes information and takes no bookings. Ask a qualified physiotherapist to examine your own case before acting on anything written here.

Guidance for patients and carers How this page was reviewed