Bone on bone: what a Grade 4 diagnosis actually means

Where the phrase comes from

The grading system behind it dates to 1957, when Kellgren and Lawrence published a way of scoring osteoarthritis on plain X-ray. It runs from Grade 0, no features, to Grade 4, defined by large osteophytes, marked narrowing of the joint space, severe sclerosis of the bone, and definite deformity of the bone ends.

Bone on bone is the everyday shorthand for a Grade 4 film: the space between the two bones, which on an X-ray is really the shadow cast by cartilage, appears to have gone.

That is a description of a two-dimensional image taken at one moment. It is useful, and it is not a prognosis, a treatment plan, or a statement about how much pain you should be in.

The grade describes the film, not the pain

This is the part that surprises people most, and it is well documented. A systematic review by Bedson and Croft examined how closely radiographic knee osteoarthritis and knee pain actually track each other. Among people reporting knee pain, radiographic osteoarthritis was present in between 15 and 76 percent of cases depending on the study. Among people with radiographic osteoarthritis, knee pain was reported in between 15 and 81 percent.

The relationship is loose in both directions. There are people walking around with Grade 4 films and manageable symptoms, and people in significant pain whose X-rays look mild.

So a severe grade does not by itself mean you must act now, nor that treatment is warranted because the picture looks dramatic. What you can do, what hurts, and what you want to get back to are the things worth measuring.

What the label leaves out

Grade 4 is a single number standing in for a joint. Two knees carrying the same grade can be quite different mechanical problems, and the differences are what a physician evaluating you for anything should be looking at:

• Which compartment is affected, and whether the rest of the joint is relatively intact
• Overall limb alignment, and whether it is loading one side of the knee disproportionately
• Ligament stability
• Meniscal status, including previous meniscal surgery
• How much active inflammation is present
• The condition of the bone beneath the cartilage
• Previous operations and what they changed
• Metabolic and inflammatory health, weight, and activity level
• What you actually want the knee to do afterwards

What the research does, and does not, show for severe disease

The evidence picture for severe osteoarthritis is not the same as the picture for mild and moderate disease, and this is where honesty matters more than enthusiasm.

The most rigorous trial in this area to date is MILES, published in Nature Medicine in 2023. It randomized 480 patients with Kellgren-Lawrence grades II through IV, meaning it did include people with severe disease, which most studies do not. Three cell-based treatments were compared against a corticosteroid injection: autologous bone marrow aspirate concentrate, autologous stromal vascular fraction, and allogeneic umbilical cord tissue mesenchymal stromal cells.

At twelve months, none of the three outperformed the others, and none outperformed the corticosteroid injection on pain or on the KOOS outcome score. That is not a marginal finding in a small study. It is a well-powered phase 3 trial reporting that expensive cellular treatments did no better than a cheap, widely available one.

Set against that, several systematic reviews do report improvement in pain and function after mesenchymal stromal cell treatment, and those are real findings. But the trials they pool largely enrolled mild to moderate disease. Grade 4 patients are frequently excluded outright, or included in numbers too small to analyse separately.

So the accurate summary is narrower than the marketing suggests: the evidence speaking directly to Grade 4 knees is thin, and the best-powered trial that did include them found no advantage over a standard injection at one year.

Why a disappointing trial result is worth your attention

Clinic websites tend to cite the encouraging studies. A field's honesty shows in how it handles the discouraging ones. This does not prove nothing works: a null result at twelve months does not rule out benefit in a subgroup, at a different dose, or measured a different way, and combinations of cells with biologic scaffolds remain an active area of study.

It does mean that anyone describing dramatic, reliable results for bone-on-bone knees is running ahead of the published evidence. If a clinic presents only the favourable literature and has no account of MILES, that tells you something about how they will describe your own chances.

Separating four different outcomes

Much of the confusion in this field comes from treating four quite different goals as one. They are not equally supported, and they are worth naming separately before any consultation.

• Pain reduction. The most reported outcome and the best supported, though MILES found no advantage over corticosteroid.
• Improved function and mobility. Usually reported alongside pain, with the same caveats.
• Delaying joint replacement. Plausible if pain and function improve, but rarely measured directly over the years that would settle it.
• Structural cartilage regeneration. The claim people most want to hear, and the least supported in severe disease. No one can promise you a rebuilt joint.

When replacement is the better answer

It is worth being even-handed about the operation people are usually trying to avoid. A systematic review covering 208 studies and more than 95,000 patients found a median satisfaction rate of around 89 percent after total knee replacement, with roughly one patient in five reporting dissatisfaction. That is a strong result by the standards of major surgery, and it still leaves a meaningful minority unhappy.

For a younger patient there is a real argument for delay, because implants have a finite lifespan and revision surgery is more complex than the first operation. But delay is only worth having if the intervening years are better. Trading a replacement you would have been satisfied with for several years of unchanged pain is not a win.

Severe deformity, major instability, or mechanical collapse can make replacement the appropriate treatment regardless of what regenerative medicine may eventually offer. Choosing it is not a failure.

The regulatory picture, plainly

The FDA's consumer alert on regenerative medicine products is direct about the current position. The only FDA-approved stem cell products are derived from umbilical cord blood, and they are approved for disorders affecting the production of blood. Nothing is approved for orthopedic use, and no exosome product has been approved.

That is why a good deal of this care happens outside the United States, and it is a fact to understand rather than a point in anyone's favour. There is no approval to point at and no domestic regulatory review standing behind the treatment, which puts more weight on the physician in front of you and on the questions you are willing to ask. The burden of evaluating a claim sits with you, and that is better acknowledged than glossed over.

Questions worth asking when you have been told bone on bone

These are specific to severe disease and candidacy, and they are the ones that turn a label back into information you can use. Ask them of any physician, in any country.

• Which compartment is involved, and how does the rest of the joint look?
• Is my alignment contributing to this, and would that need to be addressed for any treatment to hold?
• Are you evaluating me as a Grade 4 knee, or as this particular joint?
• What specifically would you expect to change, and by roughly how much?
• What does the evidence say for someone at my grade, including the trials that did not show a benefit?
• What happens if I do nothing for six months?
• What would make you tell me I am not a candidate?
• If this does not work, what does the follow-up plan look like once I am home?

The bottom line

Bone on bone is a description of an X-ray. It is not a sentence, and it is not a green light either.

The useful move after that appointment is not to search for the clinic with the most confident website. It is to convert a one-word label back into specifics: which compartment, what alignment, what stability, how much inflammation, and what you want this knee to do in two years. With those, a physician can tell you something meaningful, including that the honest answer may be a replacement, or may be to wait.

Our part is coordinating access to that conversation and being straight about what the research supports. Deciding what happens to your knee belongs to you and a licensed physician who has examined it.