OSTEOARTHRITIS · CHAPTER 12
New Treatments for Osteoarthritis: What Does “Promising” Really Mean?
Understanding new research without confusing pain relief, changes on a scan and rebuilding a joint.
On this page
1. There is genuine progress—and more than one kind of success
A treatment can help you sleep, walk or use your hands more comfortably without rebuilding cartilage. Another treatment can change a measurement on a scan without yet making daily life easier. Both findings deserve an accurate explanation.
Promising does not mean useless. It also does not mean proven for your joint. This chapter helps you ask what a study actually showed, what it did not show and whether its findings apply to the option being offered to you.
The examples concern research, not instructions to start a medicine or change an existing dose. Different joints and different forms of osteoarthritis may need different approaches.
2. Less pain, structural change and regeneration are different claims
| What you hear | The useful question |
|---|---|
| “People felt better.” | Better than which comparison group, and for how long? |
| “The cartilage measurement improved.” | Did function improve too? Were harms and missing participants considered? |
| “The blood marker changed.” | Does this marker reliably tell us that the person benefits? |
| “It regenerates the joint.” | What tissue was restored, and was that demonstrated in treated people—not only in a laboratory? |
A scan or blood measurement may help researchers, but it is not automatically a substitute for outcomes that matter to you. Regulators distinguish clinical outcomes from measurements intended to stand in for them. [3]
An illustrative example is sprifermin: a knee trial found a difference in cartilage thickness, without a clear overall difference in the measured pain/function score. That makes the structural result interesting—not proof that the knee was cured. [4]
3. What kinds of treatment are being investigated?
Researchers are studying ways to protect or rebuild cartilage, change bone remodeling, reduce local inflammation, deliver genes that produce a protective protein, use carefully defined cell preparations, and address metabolic factors. These are separate approaches—not different names for the same injection. [4, 7, 8, 15, 17, 23]
For example, a denosumab trial in erosive hand osteoarthritis found an effect on radiographic damage. It did not establish the same benefit for knee or hip OA. The research dosing also differs from usual osteoporosis treatment. [8]
Do not change or delay a prescribed denosumab injection because of this research. Stopping or delaying it can increase fracture risk, and advanced kidney disease creates important calcium-safety concerns. [9]
4. Two 2026 headlines need careful reading
Lorecivivint: on September 8, its developer reported that the UK regulator had accepted an application for assessment. This was a review milestone, not approval. Published major trials had missed their main pain endpoints; later subgroup and extension findings are a different part of the evidence. [10, 11, 13]
PCRX-201 gene therapy: a paper published in 2026 reported an early study of 72 people followed for up to two years. Symptoms improved among participants with follow-up, but there was no placebo group and 33 people discontinued before the 104-week assessment. This is a reason for further controlled research, not a reliable promise of two years of benefit from one injection. [15]
A newly published paper can report observations collected years earlier. A planned longer follow-up is not a result already observed.
5. “Stem cells” does not identify one treatment
Ask what the product actually contains, how it is prepared, and whether that exact preparation was studied in people with your condition.
One placebo-controlled study of a particular cultured fat-derived cell preparation found better symptoms at six months but no between-group advantage in cartilage defects. Another large comparison found no advantage of several cell preparations over a corticosteroid injection. A newer placebo-controlled trial was also negative for its main outcome. These results cannot be reduced to “all cells work” or “every cell treatment is identical.” [17, 18, 19]
In the United States, FDA warns that marketed stem-cell and exosome products of this kind are not approved for orthopedic conditions such as OA. A study registration or a registered facility is not the same as approval of its product. Other countries require their own status check. [20]
6. Can a blood test predict my future—or choose my treatment?
Blood and MRI studies are investigating who may develop OA or progress more quickly. A study involving 200 women found a protein pattern associated with later knee OA, years before it appeared on radiographs. That is promising research, not certainty about any one person’s future. [27]
Predicting risk and predicting response to a drug are different jobs. A marker associated with faster progression does not automatically tell us which medicine will help. [2]
A 2026 study used blood markers to propose biological subgroups and then examined earlier trial data. It has not yet established that assigning a person to one of these groups and selecting a matched treatment improves care. [29]
Before paying for a panel, ask what decision it can reliably improve today, in which population, and what would happen if the result were uncertain.
7. Five questions to ask about a study
Who took part? A trial in knees with obesity does not automatically answer a question about erosive fingers.
What was the comparison? Improvement after treatment alone does not establish how much the treatment itself contributed.
What was the main question? A favorable extra finding does not erase an unsuccessful main outcome.
How many people remained? A long follow-up among fewer remaining participants can give an incomplete picture.
What were the harms and practical costs? Include recovery time, travel, restrictions and follow-up—not only the injection price.
These are reading prompts, not a score. They help turn a headline into a useful discussion.
8. Feeling better is important—but safety remains a separate question
In a knee-OA trial involving people with obesity, semaglutide improved weight and pain outcomes. The study did not demonstrate cartilage regeneration or establish exactly how much pain improvement came from weight loss versus other effects. [23]
Where current guidance fits
The 2026 ACR guidance conditionally supports GLP-1 medicines for weight management in people with knee OA and obesity, alongside diet and exercise. This does not mean that every weight-loss medicine is approved for OA, or that cartilage regrows. Retatrutide remains investigational. 2026 ACR summary
Retatrutide and knee pain
In two large trials, retatrutide reduced knee arthritis pain and body weight more than placebo in people with excess weight. It remains investigational. The studies do not show that cartilage regrows. Treatment benefits, side effects and suitability require separate assessment. 36, 37
The reverse caution also matters: some anti-nerve-growth-factor trials improved symptoms but raised joint-safety concerns. Less pain alone does not prove that a joint is being protected. [25]
Report new symptoms rather than assuming they are an expected part of “healing.” A proposed research treatment should include a clear contact for adverse effects and an explanation of what is known—and unknown—about its safety.
9. Considering a clinical trial
Ask about the exact study, its purpose, the possibility of receiving a comparison treatment, the visits and tests, costs, permitted usual care and arrangements for problems. Ask who will explain the results and what happens when participation ends.
Informed consent is a discussion, not merely a signature. Participation is voluntary; the research team should explain withdrawal and follow-up arrangements. Being eligible does not guarantee personal benefit. [35]
A commercial procedure described as “experimental” is not necessarily part of a monitored clinical trial. Request the study details rather than relying on the word.
10. Do not put useful current care on hold
Movement adapted to your limitations, practical aids, symptom treatment and appropriate specialist assessment can remain valuable while research progresses. Their usefulness does not depend on calling them disease-modifying. [30]
Keep the existing plan visible during a research discussion: which actions continue, which restrictions the protocol requires, and who coordinates decisions. Do not independently stop prescribed medicines or postpone an already indicated assessment because a future product is being promoted.
You can be interested in new treatment and still ask for evidence. Those positions are compatible.
11. Three examples: a better question changes the discussion
“My scan improved, so I will never need surgery.” Ask what changed and whether avoiding surgery was actually studied. A structural result does not supply that promise.
“My blood test says high risk; I need the strongest treatment.” Ask whether the test predicts progression, response to this treatment, or neither in your situation. More risk is not itself proof of more benefit. [2]
“The application was accepted, so it must be approved.” Ask for the authorization, its date and its actual indication. Beginning regulatory review is not the same event. [13]
These are invented teaching examples, not descriptions of real patients.
12. Before deciding, name the benefit you want
Choose a concrete goal: walking a particular distance, using cutlery, sleeping with fewer interruptions, or another activity that matters to you.
Then separate three statements: what has been shown, what is plausible, and what remains unknown. Record the alternatives and the follow-up plan. A decision to participate, wait, seek another opinion or decline should be informed rather than pressured.
Scientific progress deserves interest. Your treatment decision deserves precision.
13. My research-treatment discussion record
Use the original, unscored sheet below to prepare questions and record an agreed decision. It does not identify the best treatment, predict your outcome or generate a prescription. Keep symptom, scan and biomarker findings separate.
My research-treatment discussion
Your notes stay in this open page only. They are not transmitted or saved after you close or reload it. Print the plan to keep a copy. Record actions already agreed with your treating team; this form does not prescribe treatment. Changing language does not translate your notes.
An original, unscored discussion record—not trial enrollment, a diagnostic test or a prescription.
Sources
- FDA–NIH BEST. Understanding Prognostic versus Predictive Biomarkers.
- FDA. Surrogate Endpoint Resources for Drug and Biologic Development.
- Hochberg et al. Effect of intra-articular sprifermin vs placebo on cartilage thickness: FORWARD randomized trial. JAMA. 2019.
- Schnitzer et al. S201086/GLPG1972 in ROCCELLA: phase 2 randomized trial. Osteoarthritis Cartilage. 2023.
- Wittoek et al. RANKL blockade for erosive hand osteoarthritis: randomized placebo-controlled phase 2a trial. Nat Med. 2024.
- FDA. Denosumab: severe hypocalcemia in advanced chronic kidney disease and precautions around discontinuation. January 19, 2024.
- Yazici et al. Lorecivivint OA-10: phase 3, 28-week placebo-controlled trial. Clin Exp Rheumatol. 2025.
- Yazici et al. Lorecivivint OA-11: phase 3, 56-week placebo-controlled trial. Clin Exp Rheumatol. 2025.
- Biosplice. UK MHRA accepts lorecivivint application and begins assessment. September 8, 2026. Sponsor announcement, not an approval or independent efficacy assessment.
- Cohen et al. PCRX-201: safety, biodistribution and exploratory clinical outcomes, two-year phase 1 results. Ann Rheum Dis. Online August 4, 2026. doi:10.1016/j.ard.2026.06.034.
- Kim et al. Autologous culture-expanded adipose-derived mesenchymal cells: phase III placebo-controlled knee-OA trial. Am J Sports Med. 2023.
- Mautner et al. Cell-based versus corticosteroid injections for knee OA: randomized phase 2/3 trial. Nat Med. 2023; corrected December 2023.
- Pers et al. ADIPOA2: adipose-derived mesenchymal stromal cells versus placebo. Ann Rheum Dis. 2025. doi:10.1016/j.ard.2025.07.026.
- FDA. Important Patient and Consumer Information About Regenerative Medicine Therapies.
- Bliddal et al. Once-weekly semaglutide in persons with obesity and knee OA: STEP 9. N Engl J Med. 2024.
- Hochberg et al. Long-term safety and efficacy of tanezumab versus NSAIDs for hip or knee OA. Arthritis Rheumatol. 2021.
- Kraus and colleagues. An OA pathophysiological continuum revealed by molecular biomarkers. Sci Adv. 2024.
- Hannani et al. Therapeutic subtypes of knee OA: differential treatment effects among predicted endotypes in past trials. Arthritis Res Ther. May 7, 2026.
- NIAMS. Osteoarthritis: diagnosis, treatment and steps to take.
- NIH. Clinical Research Trials and You: the basics.
- Conaghan et al.. TRIUMPH 4 at 68 weeks. ACR 2026 abstract 0871.
- Hunter et al.. TRIUMPH 1 OA at 80 weeks. ACR 2026 abstract 2573; efficacy estimand.
Educational information from the author’s personal site. This is not a clinic portal, medical advice for an individual, or a way to send a message to your treating team. Private expert services, when offered, are paid directly and are not billed to insurance.
Back to top