Personalized gait correction eases knee osteoarthritis and slows cartilage wear

A randomized, placebo-controlled study was published in The Lancet Rheumatology: people with medial knee osteoarthritis had their foot angles individually “tuned” when walking (slightly “toes in” or “out” by 5–10°). After a year, this individualized step correction resulted in pain reduction comparable to painkillers and less deterioration in cartilage health indicators on MRI, compared to “sham” training without changing technique.

Background

• What we are trying to fix. In medial OA, the "inner" part of the joint is overloaded. The biomechanical surrogate for this load is the knee adduction moment (KAM): the higher it is during walking, the greater the mechanical stress on the medial compartment. The idea of gating therapy is to shift the load vector so as to reduce the KAM during the step.
• Why the foot rotation angle (FPA)? A small "toe-in" or "toe-out" (usually 5-10°) can significantly reduce KAM; but the "working" direction and magnitude differ from person to person, and for some patients, standardized instructions even worsen biomechanics. Hence the focus on personalization: selecting the FPA for a specific person.
• What came before this on gate training. A review and early RCTs showed that step modifications do reduce KAM and can reduce pain, but the effect was “blurred” due to heterogeneous protocols and lack of personalization; models have emerged that predict which step modification will work for a given patient, based on minimal clinical data — a step toward moving the technique from the lab to practice.
• Alternatives with medial unloading and their limits.
  • Lateral wedges/insoles: Meta-analyses show that there is often no clinically significant effect on pain compared to neutral insoles.
  • General exercise/strength: beneficial for function and symptoms, but does not by itself guarantee reduction in KAM. Hence the interest in targeted biomechanical interventions.
• Safety and "redistribution" of loads. When changing FPA, it is important not to "shift" the problem to other joints: studies specifically checked whether moments in the hip joint were increasing - no significant deterioration was found, but monitoring is mandatory.
• Why the current work is a step forward. This is one of the first placebo-controlled, long-term studies where personalized step correction was assessed not only by pain, but also by structural MRI markers of cartilage - that is, they checked whether unloading affects the "health" of the tissue. This design answers the main reproach to gate therapy: "biomechanics are more beautiful, but it is not easier for the cartilage." (Context: in the current OARSI guidelines, the basis remains education, exercise, weight loss, and targeted unloading interventions are options "as indicated"; personalization gives a chance to increase their effectiveness.)

What did they do?

• We recruited 68 adults with confirmed medial compartment knee osteoarthritis and pain ≥3/10.
• At the start, each participant underwent a gait analysis: a treadmill with pressure sensors + a video marker system; a computer model calculated which modification of the foot rotation angle (toe-in / toe-out by 5° or 10°) best reduces the load on the “inner” part of the knee.
• Next comes randomization:
  • Intervention - trained your "best" angle, chosen according to the model;
  • Sham control - given a "prescription" equal to the usual foot angle (i.e. no real change).
    Both shoulders underwent 6 weekly sessions with biofeedback (vibration sensors on the shin) and then trained independently ~20 min/day. Repeated assessment of pain and MRI biomarkers of cartilage microstructure - after 12 months.

Main results

• Pain: In the “personalized” group, the drop was ≈2.5 points out of 10, which is comparable to the effect of over-the-counter analgesics; in the sham group, it was just over 1 point.
• Cartilage on MRI: intervention showed slower degradation of markers related to cartilage health compared to control. (We are talking about quantitative MRI markers of microstructure, not just “pretty pictures.”)
• Biomechanics: personal adjustment actually reduced the peak load on the medial compartment (on average -4%), while in the control group the load, on the contrary, slightly increased (+>3%). The participants were able to maintain the new angle with an accuracy of ~1°.
• This is the first placebo-controlled study to show long-term clinical benefit of a biomechanical intervention in knee OA - not only in terms of pain but also in terms of structural cartilage markers.

Why is this important?

Standard treatment options for knee OA include painkillers, exercise therapy, weight loss, and, if it progresses, arthroplasty. Biomechanical approaches that reduce medial joint stress through walking technique are attractive, but have so far lacked evidence from rigorous RCTs. New work shows that personalizing foot angles, rather than giving everyone the same advice, has a more consistent effect and is even visible on MRI.

How does this work

With OA, the "inner" part of the knee is often overloaded. A small (almost imperceptible to the eye) turn of the foot changes the force vector and redistributes the load, unloading the vulnerable sector of the cartilage. But what kind of turn is needed is strictly individual; the universal instruction "toes inward for everyone" can even worsen the load in some people. Therefore, gait modeling and choosing a personal angle are the key to success.

Restrictions

• Size and duration. 68 people and a 12-month time horizon are enough to see a signal, but not enough to draw conclusions about the impact on “hard” outcomes (surgeries, long-term exacerbations). Larger multicenter RCTs are needed.
• Laboratory setup. Personalization was done in a special laboratory with expensive systems. Although the authors are already testing simplified methods (smartphone video, "smart" insoles/sneakers), transferring to a regular clinic is a separate task.

What does it mean "for tomorrow"

This isn’t a one-size-fits-all “turn your toes in and it’s all over” hack. But the study does support the idea that personalized gait training could be an adjunct to therapy for early/moderate knee OA — as a way to reduce pain and preserve cartilage without drugs or devices. Keep an eye out for new protocols: The authors are already reporting on the development of more accessible personalization methods for real-world practice.

Source: Abstract of the article in The Lancet Rheumatology (August 12, 2025) and NYU/Utah/Stanford press releases/news with key figures and study design. DOI: 10.1016/S2665-9913(25)00151-1.