Back Pain Doctor
Knee Osteoarthritis
Written by Dr Joshua Hatch, Musculoskeletal GP, RACGP Fellow, Diploma in Musculoskeletal Medicine (FSEM UK).
Osteoarthritis of the knee involves gradual changes to the cartilage and other structures within the joint. It is very common with age, and the degree of change seen on imaging does not always match how much pain a person has.
The cornerstone of management is keeping the muscles around the knee strong, maintaining activity and a healthy weight, and using simple pain measures as needed. These approaches have strong evidence and remain important regardless of any other treatment.
Where pain persists despite these measures, there are additional options to consider, including injection treatments such as PRP, which has been studied specifically for knee osteoarthritis. Assessment gives a clear, evidence-based view of what each option can realistically offer for the knee, and when an orthopaedic opinion is appropriate.
Common symptoms
- Knee pain that worsens with activity and eases with rest
- Stiffness, especially after sitting or first thing in the morning
- Swelling, grinding or a feeling of the knee "giving way"
- Reduced ability to walk, climb stairs or squat
Commonly seen in
Evidence for twelve treatment options in knee osteoarthritis
This is not a fixed ladder or a package of treatments. Every option is assessed independently against this diagnosis, including treatments with limited direct research that may still be reasonable for a clearly defined tissue target. The evidence level, likely benefit, risks, cost and alternatives are weighed at consultation.
Foundation: diagnosis, education and rehabilitation
The part of care with the broadest relevance. For many people it is enough on its own; other options are judged against it, not automatically added after it.
Diagnosis, education and progressive rehabilitation Working out what is driving the pain, explaining what your imaging does and does not mean, and building a graded loading plan you can actually follow. Core care
What it involves
Assessment to identify the likely pain source, an explanation of what the diagnosis means, adjustment of the loads that aggravate it, and a progressive strength and conditioning plan, with weight management where that is clinically relevant.
What the evidence shows
Exercise is the core treatment, and it performs as well as the tablets. A network meta-analysis of 152 randomised trials and 17,431 patients found exercise gave the same pain relief and functional improvement as oral anti-inflammatories and paracetamol, with a far better safety profile (Weng, Br J Sports Med 2023). EULAR (Moseng, Ann Rheum Dis 2024) and the American College of Rheumatology (Kolasinski, Arthritis Rheumatol 2020) both make exercise, education and weight management the core of non-pharmacological care. An individual-participant meta-analysis of 31 trials and 4,241 patients is clear about the size of it: the average effect is small and of questionable clinical importance, and those with more pain and worse function at the start benefit most (Holden, Lancet Rheumatol 2023).
Pairing reviewed 2026-09-02.
Where it fits
Worth being straight about: the average effect is modest, and it is larger for people who are struggling more to begin with. What it does have over the alternatives is that it is safe to keep doing, and it works on the things a joint injection cannot: strength, capacity and weight.
Muscle and symptom-modifying options
Useful only when the examination identifies the tissue they target or when short-term pain control creates a useful rehabilitation window.
Wet needling and trigger point injection (trigger point injection) A fine needle into a tender band of muscle, with a small volume of dilute glucose solution injected through it. Dry needling evidence
What it involves
The needle is placed into the tender muscle band and a small volume of 5% glucose with a very dilute local anaesthetic (0.02-0.08%) is injected. Often done in the same visit as your assessment, and usable where the tender area sits deeper. Dry needling (the same technique with nothing injected) is not offered here.
What the evidence shows
The needling evidence at this joint is substantial, but all of it is dry needling rather than the injected form this row describes. A meta-analysis of 20 randomised trials in 1,234 people with knee osteoarthritis or patellofemoral pain found that needling myofascial trigger points reduced pain (weighted mean difference -1.25, 95% CI -1.58 to -0.92) and improved the WOMAC function score (-6.59, -8.88 to -4.29) at moderate certainty, with the authors cautioning that blinding was poor and follow-up short (Hu, PLoS One 2026). Adding trigger point dry needling to a physiotherapy programme in 40 people improved pain by a further 1.96 points and the WOMAC score by 11.86 at four weeks and three months (Morshedlou, Physiother Theory Pract 2025), and needling of the popliteus alongside exercise in 38 people held an advantage in pain and function at six months over sham needling with the same exercise (Agost-González, Sci Rep 2025). For the injected form the base is empty: a search of trigger point injection trials at the knee, quadriceps and calf returned one record, and it tested a vibrating device for making the injection more comfortable rather than whether the injection helps knee osteoarthritis (Moon, Arch Phys Med Rehabil 2019).
Pairing reviewed 2026-09-19.
Where it fits
Directed at a tender band in the quadriceps, calf or popliteus that reproduces part of your knee pain, alongside the loading programme. It does not treat the joint surface, and the trials behind it used a needle without an injectate.
Ultrasound-guided injections Guidance means the needle can be seen reaching the target. What goes into the needle depends on the diagnosis. Short-term relief
What it involves
The injection is performed under ultrasound so the needle can be watched into the target structure. What is injected is decided by the diagnosis, and corticosteroid is used sparingly, for the reasons set out below.
What the evidence shows
At the knee this is the best-tested injection there is, and the benefit is real but brief. A Cochrane review of 27 trials in 1,767 people found a corticosteroid injection reduced pain by about 1 point out of 10 more than a dummy injection or no injection (SMD -0.40, 95% CI -0.58 to -0.22; number needed to treat 8), with the effect moderate at one to two weeks, small by 13 weeks and absent at 26, and the evidence graded low quality (Jüni, Cochrane Database Syst Rev 2015). A later meta-analysis of 11 placebo-controlled trials in 842 people found the advantage over placebo was large enough to matter only in the first six weeks (Bensa, Knee Surg Sports Traumatol Arthrosc 2024). Repeating it does not help: in 140 people given triamcinolone or saline every three months for two years, pain was no different and the corticosteroid group lost more cartilage (-0.21 mm against -0.10 mm) (McAlindon, JAMA 2017). In 156 people randomised to a corticosteroid injection or to physiotherapy, the physiotherapy group had less pain and disability at one year (WOMAC 37.0 against 55.8 out of 240) (Deyle, N Engl J Med 2020). Guidance itself has been tested here: across 13 studies, ultrasound-guided knee injections were accurately placed 95.4% of the time against 82.0% by landmarks, and all but one of the studies that measured outcome favoured guidance (Kasitinon, Am J Phys Med Rehabil 2026). The American College of Rheumatology strongly recommends corticosteroid injection for knee osteoarthritis (Kolasinski, Arthritis Rheumatol 2020).
Pairing reviewed 2026-10-02.
Where it fits
Most useful as a short bridge when a flare of pain is stopping you from doing the strengthening work that has the longer-lasting evidence. Not a treatment to repeat on a schedule: see the note on cortisone.
V-ACTOR vibration therapy (vibration therapy) Low-frequency vibration applied over tender muscle. Not a shockwave, despite sharing the machine. Plausible: no studies
What it involves
A vibration handpiece is moved over the tender muscle and fascia for a few minutes. It runs from the same STORZ MASTERPULS icon console as radial shockwave but is a separate handpiece delivering a different, gentler form of energy. No needles, no anaesthetic, no downtime.
What the evidence shows
No published clinical trial of this device in this condition. It is a vibration handpiece delivered from the same console as radial shockwave, and the rationale is myofascial rather than tendon- or bone-directed.
Pairing reviewed 2026-08-30.
Where it fits
A comfort measure used within a session where muscular tightness is limiting movement. Never the reason for the appointment, and not a substitute for the treatments that do have evidence.
Energy-based treatment options
Radial pressure wave, focused shockwave and EMTT have different targets and different evidence. They are assessed separately rather than treated as one category.
Radial or focused shockwave? Read the overviewRadial shockwave (radial pressure wave) Broad treatment across tender muscle and fascia, close to the surface. Focused device did better
What it involves
A handpiece is moved over the tender tissue while ballistic pressure pulses are delivered, using a transmitter head chosen for the tissue, including heads shaped for fascia and for the muscle either side of the spine. The pulse is strongest at the skin and weakens with depth, so this is mainly a myofascial treatment, though it can reach deeper in some areas. No anaesthetic, no downtime, usually a course of sessions about a week apart.
What the evidence shows
Radial shockwave has been tested against a sham or placebo at this joint in four randomised trials, and three were positive. In 104 people all doing the same physiotherapy and home exercise, three weekly radial sessions gave better pain, WOMAC, walking and strength scores than a sham device at the end of treatment and at one and three months (Uysal, Int Orthop 2020). Eighty-nine people randomised across four radial doses or a very-low-dose placebo had less pain and a better WOMAC score at four weeks with every treated dose, by about 2 to 3.6 points out of 10 for pain, and the higher energy did more (Zhang, J Rehabil Med 2021). A smaller unblinded trial of 45, all doing strengthening, found pain and function improved with low- and medium-energy radial treatment and not with sham (Hammam, J Taibah Univ Med Sci 2020). The contrary trial is real: in 105 women with severe, long-standing knee arthritis, three low-energy radial sessions were no better than placebo for pain on movement at three months, the main outcome, with a difference only in WOMAC pain (Imamura, J Rehabil Med 2017). Head to head, the focused form did better: in 42 people with arthritis in both knees, pain fell by 4.5 points with focused against 2.6 with radial (Ko, Int J Environ Res Public Health 2022). The trials are single-centre with follow-up of one to three months, so nothing is known beyond that, and one review found shockwave added to an exercise program gave less than a clinically important extra benefit (French, Osteoarthr Cartil Open 2024).
Pairing reviewed 2026-10-24.
Where it fits
A reasonable option for knee arthritis pain, alongside the strength and weight work rather than instead of it, and particularly where the muscle and fascia around the knee are tender as well as the joint. Where the joint itself is the target, the focused form above has the better head-to-head result and the depth to reach it. Neither changes the joint surface, and the benefit has only been followed to three months.
Focused shockwave (ESWT) A true shock wave converging at depth: for tendon, joint, ligament and bone. Good support
What it involves
A focused system generates an acoustic shock wave that converges at a focal point set at a defined depth, concentrating the energy where the target is. This is the form used for tendon, joint, ligament and bone. Typically four to six sessions about a week apart, no anaesthetic and no formal downtime.
What the evidence shows
Knee osteoarthritis is where shockwave has its deepest trial base outside tendon problems, and the results favour it consistently. An umbrella review pooling eight meta-analyses found shockwave better than sham treatment for pain, by 2.0 points out of 10 (95% CI 1.5 to 2.5), and better for the WOMAC score, the Lequesne index and knee range of movement (Tang, Int J Surg 2024). A meta-analysis of 12 randomised trials found it ahead of sham for pain at short, medium and long-term follow-up and for function in the short term (Silva, Clin Rehabil 2022), and a network meta-analysis of 74 randomised trials in 3,707 people ranked shockwave first among physical treatments for WOMAC improvement beyond three months, at moderate confidence (Zhang, Front Med 2025). Energy level and treatment duration predict the response (Liao, Clin Rehabil 2019), and when the two devices were compared head to head in 42 people, the focused form reduced pain by 4.5 points against 2.6 with radial (Ko, Int J Environ Res Public Health 2022). The American Academy of Orthopaedic Surgeons guideline lists shockwave as an option that may be used for pain and function, at limited strength (Brophy, J Am Acad Orthop Surg 2022). Two limits: the trials are small and vary in dose, and where shockwave was added to an exercise programme, one review found the extra benefit over exercise alone fell short of clinically important (French, Osteoarthr Cartil Open 2024).
Pairing reviewed 2026-10-04.
Where it fits
A good option for knee arthritis pain, and the form of shockwave to choose when the joint itself is the target. It suits people whose pain is holding back the strength work and those who want to avoid or defer an injection. It reduces pain and improves function; it does not regrow cartilage or change what the X-ray shows, and it works alongside exercise and weight management rather than replacing them.
EMTT (magnetotransduction therapy) High-energy magnetic pulses applied over the area. Usually painless, no contact. Plausible: no studies
What it involves
You sit with the applicator positioned over the area for a short session. It is usually painless and needs no gel or skin contact. It is usually given as a course, commonly alongside shockwave.
What the evidence shows
There are no published randomised trials of EMTT in this condition, or in any musculoskeletal condition, that could be identified. The supporting work is laboratory work: EMTT increased tenocyte proliferation, migration and collagen expression in cultured human tendon cells (Mancini, Int J Mol Sci 2025). The nearest human evidence for a related electromagnetic therapy is negative. A randomised trial of pulsed electromagnetic field therapy added to eccentric exercise for Achilles tendinopathy found no improvement in outcome scores or tendon neovascularity over eccentric exercise alone (Ko, Orthop J Sports Med 2024).
Pairing reviewed 2026-08-30.
Where it fits
Offered only as an adjunct alongside treatment that does have evidence, never as the plan itself, and only after that gap is explained. Much of the EMTT literature is manufacturer-affiliated, which is part of why it is described this way.
Biologic and joint injection options
Each injection is assessed independently for the named diagnosis. Cost, uncertainty, likely benefit, risks and alternatives are part of the decision.
Platelet-rich plasma (PRP) Your own blood, concentrated and injected under ultrasound guidance. Beat placebo in pooled trials
What it involves
Blood is taken from your arm, spun to concentrate the platelets, and injected into the target under ultrasound guidance in the same appointment.
What the evidence shows
The best-studied use of PRP in arthritis, with pooled placebo-controlled trials in its favour and some important trials against. Pooling 18 randomised trials in 1,995 people that compared PRP with a dummy injection, PRP was statistically better on pain and function at every follow-up from one to twelve months; the gain in function (WOMAC) was larger than the smallest difference patients notice at all four time points, and the gain in pain did so at three and six months, with higher-platelet preparations doing better (Bensa, Am J Sports Med 2025). A second meta-analysis of 11 saline-controlled trials in 1,616 people was more reserved: pain was 1.1 points out of 10 lower at three months and 2.0 lower at six, with nothing at twelve, function was better at three months only, and its authors call the efficacy weak and its clinical relevance debatable (Auroux, Joint Bone Spine 2025). The largest single trial, 610 people across nine hospitals in China, found PRP better than sham saline on pain and function from six months through to five years (Chu, Knee Surg Sports Traumatol Arthrosc 2022). The best contrary evidence is Australian. In the RESTORE trial, 288 people in Sydney and Melbourne had three weekly injections of leukocyte-poor PRP or saline: at twelve months pain had fallen by 2.1 against 1.8 points out of 10 (difference -0.4, 95% CI -0.9 to 0.2) and cartilage volume on MRI was no different (Bennell, JAMA 2021). A 102-person trial in Launceston found neither one nor three injections better than saline up to twelve months (Lewis, Bone Joint J 2022). Expert bodies are split for the same reason: the RESTORE authors noted that most clinical guidelines do not recommend PRP for knee osteoarthritis, while a European sports-orthopaedic consensus regards it as a valid option for mild to moderate disease (Laver, Knee Surg Sports Traumatol Arthrosc 2024). It is not free of side effects. Across 24 trials, complications were recorded at a rate of 18.7% with PRP against 9.1% with comparison injections - about one extra for every 11 people treated - nearly all of them mild to moderate and self-limiting (Fucaloro, Arthroscopy 2025).
Pairing reviewed 2026-10-08.
Where it fits
A reasonable option for mild to moderate knee osteoarthritis that has not responded to exercise, weight management and simple pain relief. The expectation set beforehand is an improvement in pain and function over the following three to twelve months in a proportion of people - not a guarantee, and not cartilage regrowth - and the Australian trial that found no benefit over saline is part of that conversation. The knee is often more sore for a few days afterwards. Response is reviewed at about three months.
Prolotherapy (dextrose) A course of dextrose injections into and around the structure and its attachments. Good support
What it involves
Hypertonic dextrose is injected at the tender ligament and tendon attachments, and into the joint where relevant, usually as a series of treatments a few weeks apart.
What the evidence shows
A meta-analysis of 14 trials and 978 patients found dextrose prolotherapy favourable against both placebo injection and non-invasive control for pain, global function and quality of life, with dose- and time-dependent effects (Chen, Clin Rehabil 2022). A network meta-analysis of 80 trials and 6,934 patients found dextrose prolotherapy combined with physical therapy the MOST effective of ten injection-plus-therapy regimens for both pain (SMD -2.54) and global function (SMD 2.28) (Liao, Int J Mol Sci 2023). An open-label follow-up reported progressive improvement sustained to 2.5 years (Rabago, Complement Ther Med 2015).
Pairing reviewed 2026-08-30.
Where it fits
One of the better-supported uses of prolotherapy anywhere, and the evidence is strongest when it is combined with a loading programme rather than given alone.
Viscosupplementation (hyaluronic acid) A gel-like injection into an arthritic joint. Widely used, and genuinely contested. Evidence mixed
What it involves
One or a short series of hyaluronic acid injections into the joint under ultrasound guidance.
What the evidence shows
Trialled more than any other knee injection, and the best evidence is that the average effect over a dummy injection is too small to feel. A BMJ meta-analysis of 169 trials and 21,163 people based its main analysis on the 24 large placebo-controlled trials (8,997 people): hyaluronic acid reduced pain by 2 mm on a 100 mm scale more than placebo (95% CI 0.5 to 3.8), short of the smallest difference that matters to patients, and was followed by more serious adverse events (relative risk 1.49, 95% CI 1.12 to 1.98); it also found signs that small studies and selective publication had inflated earlier estimates (Pereira, BMJ 2022). Guideline bodies genuinely disagree on what to do with that. The American College of Rheumatology and Arthritis Foundation do not recommend it, while OARSI and the United States Veterans Affairs guideline do (Sabha, Osteoarthr Cartil Open 2021), and a review of 25 guidelines found the higher-quality ones inconsistent on hyaluronic acid at the knee (Gibbs, Osteoarthritis Cartilage 2023). Part of the difficulty is that injection itself often helps in trials, which makes the specific effect of the hyaluronic acid hard to isolate. A randomised trial of 104 people found hyaluronic acid combined with hypertonic dextrose outperformed hyaluronic acid with saline on physical function at six months (Hsieh, Arch Phys Med Rehabil 2022).
Pairing reviewed 2026-10-08.
Where it fits
Discussed where corticosteroid is unsuitable or has already been used repeatedly, and where you would like a non-steroid option before a surgical opinion. The small average effect and the disagreement between guidelines are stated rather than glossed over.
Nerve-targeted options
Relevant when examination identifies a superficial neuropathic pattern or a compressed peripheral nerve, not simply because pain is present.
Lyftogt perineural therapy (perineural injection therapy) A series of small dilute dextrose injections just under the skin, over tender superficial nerves. Trials against other care, no sham
What it involves
Small injections of low-concentration dextrose, usually 5%, placed just beneath the skin along tender superficial nerves. Given as a series of short sessions.
What the evidence shows
There are now trials, but none against a dummy injection. A 2026 systematic review found eight studies, 599 people, of dextrose injected around the knee and its skin nerves (periarticular-perineural injection) for knee osteoarthritis: pain and function improved within every study, the injections did as well as or better than non-injection care in the short and medium term, and against injections into the joint they did as well or better only in the short term, with injections into the joint as good or better by the medium term; no major side effects were reported, and the authors rate the studies' quality as low (Arias-Vázquez, J Integr Complement Med 2026). In a 60-person double-blind trial, perineural 5% dextrose around the knee alone did less well on pain and WOMAC at four and eight weeks than the same injection combined with an injection into the joint, though it raised pressure-pain thresholds more than the joint injection alone (Fu, J Orthop Surg Res 2024). A retrospective review of 10 patients given three monthly ultrasound-guided injections of 5% dextrose around the back of the knee reported a 52% fall in WOMAC held at twelve months, and its authors state that placebo, regression to the mean and natural fluctuation cannot be excluded (Lam, Diagnostics 2026). Without a placebo-controlled trial, how much of the improvement comes from the injection itself is not known.
Pairing reviewed 2026-11-11.
Where it fits
Can be discussed where examination finds tender, hypersensitive skin nerve territories around the knee rather than joint-line pain alone, as a way to settle pain enough to keep the strengthening and weight program going. It is not the same as an injection into the joint, which has its own evidence, and it does not change the arthritis itself.
Ultrasound-guided nerve hydrodissection Fluid placed around a nerve to free it from the surrounding tissue. Only if relevant
What it involves
Under ultrasound, fluid is placed in the plane around the nerve to separate it from the tissue that is tethering or compressing it.
What the evidence shows
Knee osteoarthritis appears in the hydrodissection literature, but as joint distension rather than nerve treatment. A scoping review of 21 included studies lists it among the conditions where ultrasound-guided hydrodissection and hydrodilatation may improve pain and function, while pooling the two techniques and noting the evidence is early and heterogeneous (Leahy, Front Pain Res 2026). The row here is hydrodissection of a nerve, a different procedure with a different target, and no trial has tested that for knee osteoarthritis.
Pairing reviewed 2026-09-08.
Where it fits
Only where a nerve is contributing separately to the knee pain. The arthritis itself is treated by the options higher up this page.
This is general information only and suitability is assessed individually. Treatments with limited or condition-dependent evidence may still be discussed, but only with clear explanation of the uncertainty, the expected benefit, risks, cost and alternatives. Red flags, progressive neurological symptoms or suspected serious pathology require a different pathway.
Every treatment above is described in full on the non-surgical treatments page, including what each one is used for and what the evidence supports. Ratings apply to the named diagnosis; an adjacent muscle, tendon, joint or nerve problem is assessed as a separate target. The method behind the ratings, and the treatments not offered, are set out on the evidence page.
Corticosteroid injection is reviewed separately for each diagnosis.
Cortisone reliably reduces pain in the short term, and that can be genuinely useful when pain is stopping you sleeping or moving enough to rehabilitate. Beyond that window the picture changes. A systematic review of 41 randomised trials and 2,672 participants found that corticosteroid injection reduced pain in the short term but that the effect reversed at intermediate and long term; in pooled analysis for lateral epicondylalgia the short-term effect was large (SMD 1.44, 95% CI 1.17 to 1.71) while no intervention was favoured at 26 weeks (-0.40, -0.67 to -0.14) and at 52 weeks (-0.31, -0.61 to -0.01) (Coombes, Lancet 2010). In a randomised trial in lateral epicondylalgia, 83 per cent of the corticosteroid group had completely recovered or much improved at one year against 96 per cent given a placebo injection, and recurrence ran at 54 per cent against 12 per cent; adding physiotherapy made no significant difference at one year (Coombes, JAMA 2013). In joints, repeated corticosteroid injections have been associated with cartilage loss. It does not repair anything.
So treatment plans are not built around it. Where an injection is the right call, the specific diagnosis and target are reviewed, what is being injected and why is discussed, the number is kept low, and it is paired with the loading work that changes the longer-term course.
When it might still be considered
- Pain severe enough to block rehabilitation
- A clear inflammatory or bursal target on imaging
- A specific short-term goal: a trip, a procedure, a return to work
- Diagnostic use, where the response shows where the pain is coming from
When a cortisone injection is worth doing, and when it is not
Other knee conditions
Pain in the same area can come from more than one source, and the label matters less than what the examination and history point to. These are the other guides for this region.
Related articles
Shorter articles that answer the questions patients ask most often about this.
- PRP for Early Knee Arthritis: Encouraging Results for Planned CareA new trial found better pain and function when PRP for early knee arthritis was given within a structured programme. The findings and their limits.
- Baker's Cyst: Does It Need Draining?Does a Baker's cyst need draining? When watching it, knee rehabilitation or aspiration helps, and when calf swelling needs urgent care in Brisbane.
- Meniscus Tear: Do You Need Surgery?Most degenerative meniscus tears start with exercise. What the trials show, and which knee symptoms need prompt medical review in Brisbane.
- Cortisone vs PRP for Knee Osteoarthritis: What the Evidence SaysHow cortisone and PRP injections compare for knee osteoarthritis: what each does, what the evidence shows, and where injections fit alongside exercise.
- Knee Osteoarthritis Flare Plan: What to Do When Pain SpikesKnee arthritis symptoms rise and fall. A practical plan for flares that protects movement without overreacting to pain.
- Knee Pain and ArthritisWhat drives knee osteoarthritis pain, and how exercise, weight loss, medicines, injections and surgery compare.
Frequently asked questions
Is exercise safe if I have knee osteoarthritis?
What about PRP for my knee?
Will I need a knee replacement?
Ready for a clearer plan for your back or musculoskeletal pain?
Book an assessment with Dr Joshua Hatch.
Your assessment focuses on understanding the likely source of your pain and the most appropriate non-surgical options for your diagnosis, with the aim of reducing pain and improving function.
You must have your own regular GP before booking. This service works alongside your GP for focused care. Booking requirements.
Book an appointment