Why physio sometimes isn't enough

Doing the exercises, attending every session, and still waking up with the same knee pain — that experience is more common than it should be, and it carries a clinical meaning worth understanding. Physiotherapy is not failing you; in many cases, it is revealing that something structural needs attention first.

NHS and Chartered Society of Physiotherapy guidance recognises 6–12 weeks of supervised physiotherapy without meaningful improvement as the threshold for seeking further assessment. Beyond that point, continuing the same programme rarely changes the outcome. The reason is straightforward: exercises strengthen muscles and improve movement patterns, but they cannot repair a torn meniscus, restore lost cartilage, or reverse advanced arthritis. Without an accurate diagnosis, the wrong problem is being treated.

NHS orthopaedic referrals from GPs are routinely rejected unless physiotherapy has already been completed and documented — so attendance matters both clinically and practically.

Seek urgent advice if any of the following apply

• The knee was injured in a fall or collision
• You cannot put weight through the leg
• The joint is hot, red, and significantly swollen (possible infection)
• The knee locks or gives way unexpectedly
• Pain is rapidly worsening rather than plateauing

Getting the diagnosis right before choosing a treatment

Persistent knee pain after a full course of physiotherapy is a clinical signal, not a personal failure — and the most productive response is not another treatment but a clearer picture of what is actually wrong.

Many cases that plateau in physiotherapy do so because an underlying structural problem was never confirmed: a partial meniscus tear, patellofemoral cartilage thinning, or osteoarthritis that has reached a grade exercises alone cannot address. These conditions show little or nothing on a standard X-ray, which captures bone density well but misses cartilage loss, soft-tissue damage, and early chondral change entirely. MRI is the appropriate next investigation — it distinguishes between the tissue types that matter most when knee pain refuses to settle.

The referral route is straightforward. A GP can direct patients either to an extended scope physiotherapist (ESP), who carries imaging authority including X-ray and MRI requests, or directly to an orthopaedic specialist. Both pathways are clinically valid; the important thing is that imaging is arranged rather than deferred.

Reading that imaging precisely also matters. Subtle cartilage changes — early chondral softening, small focal defects — can be understated on a standard radiologist report. AI-assisted analysis tools such as onMRI™ are designed to interrogate these findings in more granular detail, which is where they sit naturally in the diagnostic process: not as a replacement for specialist review, but as a layer that sharpens it before treatment decisions are made.

Without a confirmed structural cause, any choice of injection, procedure, or surgery remains an informed guess.

Conservative steps that go beyond exercises

Whilst diagnostic steps are being arranged, several evidence-based measures can reduce pain and slow deterioration — most can be started without a prescription.

Weight and load. For osteoarthritis particularly, even modest weight reduction has a measurable effect on joint loading and systemic inflammation. A loss of 5–10% of body weight can meaningfully lower the mechanical forces transmitted through the knee with every step.

Activity substitution. High-impact exercise need not stop entirely. Swimming and cycling maintain muscle strength and cardiovascular conditioning without compressing the joint through a painful range — both are endorsed in NHS guidance for knee OA management.

Bracing and footwear. An unloader brace shifts force away from the more damaged compartment; hinged bracing adds stability where the joint is unstable. Supportive footwear and orthotics can address minor alignment issues at relatively low cost.

Analgesia. Paracetamol and NSAIDs, taken regularly at recommended doses rather than sporadically, remain first-line for symptom control. Topical capsaicin or lidocaine preparations offer a lower-risk adjunct with fewer systemic effects. Any changes to ongoing medication are best discussed with the GP at the point of referral.

Long-term opioids (codeine, morphine) are generally avoided: tolerance can develop within two to four weeks, and there is a documented risk of opioids paradoxically heightening pain sensitivity over time.

These steps are most usefully thought of as a parallel track — reducing load and managing symptoms whilst the structural cause is being identified — not a substitute for diagnosis.

Injection options: from anti-inflammatory to biological

Not all injections work the same way — and understanding the difference between them matters when the choice has to be made.

Corticosteroids reduce inflammation quickly, often settling acute flare pain within days. The effect is temporary, typically lasting weeks to a few months, and repeated use over time carries a risk of local tissue damage. For managing a severe flare before further assessment, they remain a reasonable tool; as a long-term strategy, they are not.

Hyaluronic acid (viscosupplementation) improves joint fluid quality rather than suppressing inflammation directly. Cochrane reviews and meta-analyses support its use in mild-to-moderate OA, where restoring lubrication reduces friction and discomfort.

Platelet-rich plasma (PRP) concentrates the patient's own growth factors. A 1-year RCT in 160 patients unresponsive to conventional treatment found PRP produced significantly better WOMAC pain and SF-36 quality-of-life scores than hyaluronic acid (P<0.001). Preparation protocols vary considerably between clinics, which affects how directly evidence from any one study translates to clinical practice — a point worth raising when comparing options.

BMAC (bone marrow aspirate concentrate) delivers a denser biological payload, including mesenchymal stem cells. A 4-year prospective study in 37 knees with severe OA (KL grade III–IV) recorded a 95% success rate — IKDC rising from 56 to 73 and WOMAC falling from 40 to 18 — with no prostheses required. The sample is small, and results should be interpreted in that light; four-year follow-up data of this quality are nonetheless rare in this field.

Autologous adipose-derived MSCs, assessed in a Phase III RCT of 261 patients (KL grade 3), produced superior VAS pain improvement (25.2 vs 15.5, P=0.004) with MRI-confirmed cartilage change at six months — one of the more substantial controlled datasets currently available for any biological injection.

The progression from corticosteroid to biological reflects a shift in therapeutic aim: from suppressing symptoms to actively modifying the joint environment. Which point on that spectrum suits a given patient depends on OA grade, structural defect pattern, and whether enough viable tissue remains to mount a biological response — which is why imaging findings drive the conversation rather than follow it.

Regenerative and minimally invasive procedures

Beyond biological injections, a further layer of procedures sits between established injectables and surgery — suited to patients where structural damage calls for something more targeted, but where replacement remains premature.

NanoACi (Non-Arthroscopic, Needle-delivered, One-stop Autologous Chondrogenic Injection) was developed by Professor Paul Lee and delivers autologous micrograft-based chondrogenic material — including Mytocel MSK — through needle access in a single session, with no theatre and no arthroscopy required. The aim is to create a chondrogenic repair environment using the patient's own tissue and blood, rather than merely cushioning or suppressing inflammation. The component materials have published support, and the procedure has been applied clinically in suitable patients. Combined-protocol outcome data for NanoACi specifically are still being gathered; compared with more established biologics such as BMAC, this is a newer approach with an evidence base still in development — something patients considering it should weigh honestly.

ChondroFiller is a collagen-scaffold injectable designed for focal cartilage defects — where damage is localised and healthy cartilage borders remain on imaging. It provides structural scaffolding for repair and is most relevant where MRI confirms a contained lesion rather than diffuse, joint-wide deterioration.

Genicular artery embolization (GAE) approaches pain from a vascular angle: an outpatient interventional radiology procedure that selectively reduces blood flow to the abnormal micro-vasculature that develops around arthritic joints. Evidence is still emerging, and patient selection depends closely on OA pattern and vascular assessment rather than general criteria.

What determines whether any of these is appropriate is a precise structural diagnosis — current imaging, OA grade, and defect pattern reviewed by a specialist — not the severity of symptoms alone.

Where surgery fits — and what to confirm first

Surgery has a clear and legitimate role — when it matches what imaging has confirmed.

Arthroscopic surgery remains the appropriate route for confirmed structural problems: a meniscus tear with mechanical symptoms, loose bodies causing locking, or damage too localised for injection-based treatment alone. Osteotomy can be the correct answer where joint malalignment is driving asymmetric load and accelerating deterioration on one compartment. In both cases, the indication is structural, not symptomatic — imaging drives the decision.

Total knee replacement is different in kind. It is permanent and irreversible: once a joint is replaced, biological and regenerative options are no longer available to that knee. That is not an argument against replacement — for joints where the cartilage loss is end-stage and no viable tissue remains, it delivers real and lasting relief. The concern is sequencing. A significant number of patients are referred for replacement before being assessed for joint-preserving treatments — biological injections, targeted minimally invasive procedures, surgical repair short of replacement — that might have preserved the joint for years longer.

The practical question to ask before any surgical pathway is committed to: has an imaging-confirmed diagnosis been made, and have biological and minimally invasive options been formally reviewed for suitability? Those two steps, in that order, are where the decision should begin — regardless of which clinic or specialist is involved.

1. [1] Clinical Efficacy and Safety of Autologous Adipose-Derived MSCs for Knee OA: Phase III RCT. (2023). https://doi.org/10.1177/03635465231179223 https://doi.org/10.1177/03635465231179223
2. [2] Knee Osteoarthritis Injection Choices: PRP vs Hyaluronic Acid (1-year RCT). (2015). https://doi.org/10.4137/CMAMD.S17894 https://doi.org/10.4137/CMAMD.S17894
3. [3] Intra-articular BMAC (mesenchymal stem cells) in KL grade III/IV knee OA: 4-year results of 37 knees. (2024). https://doi.org/10.1038/s41598-024-51410-2 https://doi.org/10.1038/s41598-024-51410-2
4. [4] Intra-Articular Injection of Bone Marrow–Derived MSCs in Knee Osteoarthritis (RCT). (2025). https://doi.org/10.1177/09636897241303275 https://doi.org/10.1177/09636897241303275