Why conservative treatment comes first — and what six weeks means

Being referred to physiotherapy before anything else is not a delay — it reflects how knee osteoarthritis responds best to treatment. For grades I–III, the joint still has meaningful capacity to adapt, and structured conservative care (exercise, load management, manual therapy, and weight optimisation where relevant) addresses the underlying mechanics rather than masking them. Surgery is reserved for a later stage: when conservative care has genuinely failed, symptoms are severely limiting basic activities, and imaging confirms advanced cartilage loss consistent with those symptoms.

Six weeks matters because it is the minimum window over which functional change becomes detectable and clinically meaningful. Trial data in grade II–III knee OA shows WOMAC scores can nearly halve by the six-week mark — but the figure is a signal of trajectory, not a finishing post. The six-week review is a checkpoint: to assess whether the programme is working, consider adjunct options, and plan the next phase. Some patients need longer; others plateau earlier and need a different approach.

Pain, however, does not map directly onto structural change. Imaging findings can exist without symptoms, and significant pain can persist with only modest structural involvement. Both directions matter when interpreting a six-week response.

How cartilage gets its nutrition — and why movement is the delivery mechanism

Adult articular cartilage has no blood supply. Unlike most tissues in the body, it cannot draw nutrients directly from vessels running through it — there are none. Its primary nutritional source is synovial fluid, the lubricating liquid that fills the joint space. Animal model data make the consequence of disrupting that supply clear: cartilage deprived of synovial fluid shows measurable thinning within weeks, alongside loss of key structural proteins — collagen type II and aggrecan — and elevated enzymes associated with matrix breakdown.

Movement is what keeps this supply system functioning. Each time the joint loads and unloads — during walking, cycling, or even a simple straight-leg raise — the cartilage matrix compresses fractionally, then recovers. That compression-and-release cycle works like a sponge being squeezed and released under water: fluid carrying oxygen and nutrients is drawn in on decompression; waste products are expelled on compression. Without regular movement, the exchange slows, and the cartilage is left in a comparatively nutrient-poor environment. Over time, that relative deprivation may contribute to the degenerative changes seen in prolonged inactivity and sedentary OA progression.

This mechanism underpins every form of physiotherapy exercise used in knee OA — from gentle range-of-motion work and quadriceps strengthening to aquatic therapy, where buoyancy reduces joint stress whilst preserving the loading cycle. The aim is not to regrow lost cartilage or reverse structural change; it is to maintain the fluid-exchange environment that sustains the cartilage that remains.

What happens inside the joint during each session

The most direct evidence of what exercise does to cartilage in a living knee OA joint comes from MRI. In a study of 20 adults with clinical knee OA, 3-tesla imaging taken before and immediately after a 25-minute treadmill walk showed measurable reductions in cartilage thickness — −0.088 mm at the medial femur (p=0.002) — alongside falls in T2 relaxation times across all cartilage regions. T2 relaxation reflects the hydration state and structural organisation of the collagen matrix; when it drops after loading, it signals active fluid redistribution within the tissue itself. This is real-time, in-human confirmation that the exchange process described above is not a theoretical construct.

At the cellular level, the compression sensed by the cartilage matrix is detected by integrin mechanosensors on the surface of chondrocytes, which respond by activating several repair and maintenance pathways — among them TGF-β/SMAD signalling — that drive the production of structural matrix components such as collagen and aggrecan. The cellular response to appropriate mechanical input is broadly anabolic: the chondrocyte reads correctly dosed loading as a prompt to maintain and renew the tissue around it.

That response is tightly dose-dependent, and the risk of exceeding it is real. In animal models, a single bout of overloading produces transient synovial inflammation; when overloading is repeated, the inflammation becomes prolonged. The key mediator appears to be inhibin subunit beta A (INHBA), which is upregulated in the superficial synovium under repeated excessive load and exerts catabolic effects on chondrocytes — the biochemical opposite of the therapeutic signal. The line between anabolic and catabolic is not a safety margin; it is a biological threshold.

Supervised, progressive physiotherapy is therefore not a cautious convention — it is how the joint is kept within the anabolic window. Incrementally increasing load as strength and neuromuscular control develop tracks the tissue's adaptive capacity; pushing beyond it risks shifting the cellular balance from matrix maintenance towards matrix degradation.

What the clinical evidence shows at the six-week mark

Six weeks is not an arbitrary checkpoint. The trial evidence consistently shows that clinically meaningful functional improvement — the kind that registers in daily life, not just on a score sheet — is detectable by the six-to-eight-week mark across OA severity grades I to III.

The benchmark used to judge that improvement is the minimal clinically important difference (MCID) on the WOMAC scale, which measures pain, stiffness, and physical function. Crossing it means the patient's function has genuinely changed in a way they are likely to feel, not merely shifted by the margin of measurement noise. In a 12-week RCT of 144 adults with grade II–III knee OA, WOMAC scores in the active physiotherapy arms were roughly halved by the six-week assessment — a shift that comfortably clears the MCID threshold. An 8-week pilot trial (the KneE-PAD study, KL grades 1–3) found both conventional physiotherapy and sensor-guided telerehabilitation exceeded the MCID, with mobility and quadriceps strength improving significantly in both groups.

Manual traction — a hands-on technique applied twice weekly — can accelerate this timeline further. In one study, pain and stiffness scores fell sharply within four weeks, with physical function scores following, likely because traction also reduces circulating pro-inflammatory cytokines such as IL-1β, not just offloading the joint mechanically.

The six-week review also frames a key clinical decision: whether to add an injection alongside continuing physiotherapy. A systematic review of five RCTs (552 patients) found that intra-articular injections produced faster short-term pain relief than physiotherapy alone, whereas physiotherapy delivered greater long-term functional improvement — particularly in earlier-stage disease. Combination therapy produced the strongest short-term outcomes. If function at six weeks has not crossed the MCID threshold, that evidence is directly relevant to what the clinician considers next.

When physio appears to fail — and why that judgement deserves scrutiny

Declaring physiotherapy a failure requires one condition that is frequently unmet: the patient actually completed a structured programme at a therapeutic dose. Attendance data consistently show that completion is not uniform. Gender, BMI, pain on physical function testing, prior history of knee injections, and pre-existing psychological symptoms all independently predict whether patients follow through with a structured knee OA programme. These are clinical variables, not patient failures — and they are identifiable before the first session begins.

Non-completion and treatment failure are not the same outcome, and conflating them produces incorrect escalation decisions. A patient who attended four of twelve sessions has not experienced physiotherapy; they have experienced a partial exposure to it. The escalation clock does not restart from the final attended session — it restarts from the end of a completed programme at adequate dose.

Psychosocial screening therefore belongs at the first appointment, not at the six-week review when a referral is already under consideration. Identifying barriers early — and addressing them through supported self-management, adapted delivery, or psychological input where indicated — is part of the treatment, not a preliminary to it. A programme that was never completed cannot be fairly evaluated as a treatment that was tried and found wanting.

When to escalate — red flags, review checkpoints, and the surgical threshold

Three distinct thresholds govern escalation in knee OA, and recognising which applies avoids both under-treatment and unnecessary delay.

Stop and seek urgent assessment

Certain presentations bypass the six-week physiotherapy trial entirely. Severe night pain that consistently prevents sleep, a knee that is hot, swollen, and visibly red — which raises the possibility of septic arthritis or an acute inflammatory flare — and sudden, repeated locking or instability that prevents weight-bearing all require same-day or urgent clinical assessment before a conservative programme begins. The same applies to new leg weakness, numbness, or pins-and-needles, which may indicate a neurological cause that physio is not designed to address.

The six-week checkpoint

For patients following the standard conservative pathway, the six-week mark is a trajectory review, not a pass/fail verdict. If functional gains are meaningful and progressing, the programme continues. If pain remains severe or function is still significantly impaired, this is the appropriate point to consider adjunct options — corticosteroid or hyaluronic acid injection, medication review, or onward specialist assessment — rather than continuing unchanged and reassessing at twelve weeks. A specialist review at this stage is also appropriate when the clinical picture has become less clear since the programme began.

The surgical threshold

Surgical referral requires three factors to converge: failure of conservative treatment across three to six months of structured exercise, severe functional limitation in everyday activities such as stair-climbing or short-distance walking, and radiological evidence of bone-on-bone change that is consistent with — not merely coincident with — the patient's symptoms. Imaging findings alone do not drive this decision; the structural picture must match the clinical one.

That convergence requirement matters in both directions: it protects patients from premature escalation when imaging looks alarming but symptoms are manageable, and it ensures those with genuine functional failure are not left waiting indefinitely on conservative care that has already reached its ceiling.

1. [1] Randomized Trial Comparing AI-Tailored Home Physiotherapy Versus Clinic-Based Rehab in Knee Osteoarthritis. (2025). https://doi.org/10.71000/52a9k939 https://doi.org/10.71000/52a9k939
2. [2] What patient-specific factors can potentially affect physiotherapy attendance of patients with knee OA?. (2025). https://doi.org/10.1142/S1013702525500064 https://doi.org/10.1142/S1013702525500064
3. [3] Maximizing Knee OA Treatment: A Comparative Look at Physiotherapy and Injections. (2024). https://doi.org/10.3390/jpm14111077 https://doi.org/10.3390/jpm14111077
4. [4] Wearable Sensor–Based Telerehabilitation Versus Conventional Physiotherapy in Knee OA: Insights from the KneE-PAD Pilot Study. (2025). https://doi.org/10.3390/app152412988 https://doi.org/10.3390/app152412988
5. [5] The Effect of Physiotherapy Manual Traction Techniques on the WOMAC in Knee Osteoarthritis Patients. (2025). https://doi.org/10.35451/jkf.v7i2.2466 https://doi.org/10.35451/jkf.v7i2.2466
6. [6] Exercise-Induced Changes in Knee Cartilage In Vivo: Comparing MRI Sequences. (2025). https://doi.org/10.1002/jor.70043 https://doi.org/10.1002/jor.70043
7. [7] Nutrition and degeneration of articular cartilage. (2012). https://doi.org/10.1007/s00167-012-1977-7 https://doi.org/10.1007/s00167-012-1977-7