Selecting the right primary endpoint is one of the earliest and most consequential decisions in an osteoarthritis development program. Regulatory teams must distinguish between endpoints FDA has accepted in prior approvals and those that remain investigational, because that distinction shapes trial design, labeling negotiations, and the feasibility of a given claim.
The analysis below maps the two principal claim categories FDA recognizes in osteoarthritis — symptomatic and structural — covering the specific endpoints FDA has accepted or rejected, the guidance documents that define each pathway, and the evidentiary standards that apply to each track.
Want Rhizome's help on your own question? Try it for free.
FDA-accepted endpoints in osteoarthritis: symptomatic relief versus structure modification
Osteoarthritis (OA) drug development splits cleanly into two regulatory tracks that FDA treats very differently. One track supports symptomatic claims, the signs and symptoms of the disease (pain, physical function, and the patient's global assessment). The other track supports structure-modification or disease-modification claims (slowing or reversing the structural progression of the joint). Every OA product FDA has approved to date sits in the first track. The second remains, in FDA's own words, an unmet need with no accepted endpoint. This article lays out what FDA has actually accepted for each and where the evidentiary bar sits.
The two claim categories
FDA's thinking is captured in its 2018 draft guidance, Osteoarthritis: Structural Endpoints for the Development of Drugs 191. That document is explicitly about structural endpoints for products intended to treat the underlying pathophysiology and structural progression of OA, and it states directly that it does not address symptom improvement such as pain or functional impairment. It notes that OA approvals to date have rested on patient-reported outcomes measuring pain and function, and that symptom endpoints would be handled in future guidance 4. In other words, FDA formally partitions the field: the symptomatic pathway is the established, well-trodden one, and the structural pathway is the frontier the agency was still trying to define.
Symptomatic endpoints FDA has accepted
For symptomatic claims, FDA has converged on a triad of outcomes: pain, physical function, and a patient global assessment 4. General chronic-pain and analgesic development guidance reinforces this: change in pain is the primary efficacy endpoint, ideally paired with physical function, and FDA notes that pain and function should be measured with separate validated instruments because no single instrument measures both well; it names WOMAC among the validated disease-specific measures 6. Patient global assessment (or patient global impression of change) is a standard secondary outcome 109.
The workhorse instrument across approved products is the WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index), used as a composite and by its pain, stiffness, and physical-function subscales, together with visual analog or numeric pain scales and a patient/physician global assessment. The three-domain structure (WOMAC pain, WOMAC physical function, and a patient global rating) recurs as the co-primary efficacy framework FDA applies to OA trials 113.
Approved products and their supporting endpoints:
| Product (active) | Route / setting | Primary / co-primary efficacy endpoints | Indication wording |
|---|---|---|---|
| Celebrex (celecoxib) | Oral NSAID | Patient's Assessment of Arthritis Pain (100-mm VAS), Patient's and Physician's Global Assessment of Arthritic Condition, WOMAC composite and pain/stiffness/function subscores 343536 | "treatment of the signs and symptoms of osteoarthritis" of the knee and hip 385255 |
| Voltaren / topical diclofenac | Topical NSAID | Three co-primary endpoints: WOMAC pain, WOMAC physical function, and patient global assessment (global rating of disease) 59717278 | indicated for OA of joints amenable to topical treatment, such as the knees and hands 178 |
| Cymbalta (duloxetine) | Oral SNRI | Primary: change in Brief Pain Inventory (BPI) 24-hour average pain (weekly mean); secondary: WOMAC subscales, Patient Global Impression of Improvement, Clinical Global Impression of Severity, 30% and 50% pain responder rates 8696100102103 | studied and approved for chronic musculoskeletal / osteoarthritis knee pain 9496 |
| Zilretta (triamcinolone acetonide, extended-release) | Intra-articular corticosteroid | Primary: change from baseline to Week 12 in the weekly mean Average Daily Pain intensity (0-10 NRS) vs placebo; key secondary: WOMAC function, PGIC, and area-under-effect pain analyses 137141142 | "treatment of pain of osteoarthritis in the knee" / moderate-to-severe OA knee pain 143146149 |
Two patterns are worth flagging for RA professionals. First, the indication language is consistently framed around symptoms, "signs and symptoms of osteoarthritis" or "management/treatment of pain of osteoarthritis of the knee", never a structural or disease-course claim 38143. Second, the endpoint architecture is stable across drug classes: oral NSAIDs, topical NSAIDs, SNRIs, and intra-articular corticosteroids were all cleared on the same pain-plus-function-plus-global logic, differing mainly in the specific pain instrument (WOMAC pain, VAS, BPI average pain, or a daily-pain NRS) and the trial duration.
Instruments and analyses that carry the symptomatic case
Beyond the co-primary structure, FDA relies on a recognizable toolkit for the symptomatic pathway:
- Pain measurement: WOMAC pain subscale, 100-mm VAS pain, and 0-10 numeric rating scales of average daily pain 3659137.
- Function: WOMAC physical-function subscale 59100141.
- Patient perspective: Patient Global Assessment / global rating of disease and Patient Global Impression of Change or Improvement 3486141.
- Responder analyses: proportion of patients achieving at least 30% and at least 50% reduction in pain, plus cumulative responder curves and rescue-medication use as supportive endpoints 109111. In intra-articular hyaluronic-acid and related knee-OA programs, OMERACT-OARSI responder definitions (built from pain, function, and global thresholds) also appear as prespecified outcomes 57657576.
This is the practical answer to "what does FDA accept for a symptomatic OA claim": a pain co-primary plus physical function, supported by a patient global measure and responder analyses, using validated instruments (WOMAC most prominently) over a defined treatment window.
Structure-modification and DMOAD claims: still no accepted endpoint
The structural side is where the regulatory gap sits. FDA's 2018 draft guidance frames inhibition of structural damage as an unmet need, but states that it is unclear what magnitude of change in a structural endpoint would translate into a clinically meaningful benefit, and that no structural endpoint has been used for traditional or accelerated approval in OA to date 2.
The evidentiary bar FDA describes is high and surrogate-focused. To support a structure-modifying or disease-modifying OA drug (DMOAD) claim, the evidence must give substantial confidence, from randomized controlled-trial data and/or a comprehensive understanding of the disease process and mechanism of action, that an effect on the structural endpoint will reliably predict a clinical outcome such as reduced pain, improved function, or delayed progression to end-stage disease and joint replacement 2. FDA frames the ultimate goal as avoiding or delaying joint failure and joint replacement while reducing worsening pain and function 2. Structural imaging measures such as radiographic joint space and MRI are discussed as candidate structural readouts, but the guidance does not lock in a validated joint-space-width or joint-space-narrowing threshold or an approval standard for OA 2.
The contrast with rheumatoid and psoriatic arthritis is instructive, because it shows what an accepted structural claim actually looks like when the surrogate is validated. In the RA/PsA setting, FDA has accepted a "radiographic response" claim based on the modified Total Sharp Score (mTSS), which quantifies joint space narrowing and erosions on hand and foot radiographs using the van der Heijde modification, typically over at least 12 months 28293032. FDA characterizes that as a claim of "inhibiting progression of structural damage," with the smallest detectable difference for the van der Heijde-modified Sharp score around 5 units 283032. Critically, that accepted radiographic-response precedent is an RA/PsA precedent, not an OA one 293132. FDA has not extended an analogous structure-modification or disease-modification claim to osteoarthritis; the joint space narrowing component in OA has not been established as a surrogate that reliably predicts patient benefit 2.
Current guidance landscape
Two draft guidances define the present state of play. The OA-specific structural document, Osteoarthritis: Structural Endpoints for the Development of Drugs, remains a draft dated August 2018 191, and it deliberately leaves symptom endpoints to separate guidance while confirming that no structural endpoint has yet supported an OA approval 24. On the symptomatic side, FDA's chronic-pain endpoint expectations now flow through its broader analgesic guidance, including the May 2026 draft Development of Non-Opioid Analgesics for Chronic Pain 185 (with a companion acute-pain draft 186), which articulate the pain-plus-function primary structure, validated instruments including WOMAC, patient global measures, and 30%/50% responder analyses that OA symptomatic programs follow 6109111.
Practical takeaways for OA programs
- A symptomatic OA claim is the established, approvable path. Build the pivotal package around a pain co-primary plus physical function, add a patient global assessment, and prespecify responder analyses; WOMAC (composite and subscales) is the default validated instrument, and the indication will read as "signs and symptoms" or "management of pain" 3459113.
- The specific pain instrument is flexible (WOMAC pain, VAS, BPI average pain, or a daily-pain NRS have all been accepted), but pain and function should be captured with separate validated instruments 6137.
- A structure-modification / DMOAD claim has no accepted endpoint in OA. A sponsor pursuing one should expect to justify, prospectively and with FDA, why a structural readout (radiographic joint space, MRI) will reliably predict clinical benefit, an argument FDA has accepted in RA/PsA via the mTSS radiographic-response claim but not, so far, in OA 22832.
- Because the OA structural guidance is still a 2018 draft and the analgesic guidance was refreshed in 2026, sponsors on either track should confirm current expectations directly with the review division before finalizing endpoints 185191.