Omega-3 has the best evidence of any joint supplement, and it is for rheumatoid arthritis rather than osteoarthritis, at around 2.7g of combined EPA and DHA daily. Curcumin comes second for knee osteoarthritis, with consistent positive trials and low certainty in the underlying reviews. Undenatured type II collagen is third on a small number of industry-funded trials. Glucosamine and chondroitin rank fourth because NICE tells clinicians not to offer glucosamine, and MSM fifth because its meta-analysis found no significant benefit.
Supplements marketed for building muscle answer a different question, and the evidence there is assessed in PhysioHub’s guide to supplements for muscle growth.
Ranked by the quality and consistency of the trial evidence, with the condition each one actually applies to.
| Supplement | Applies to | Evidence quality | Effective dose in trials | Time to judge |
|---|---|---|---|---|
| 1. Omega-3 | Rheumatoid arthritis mainly | Moderate, consistent | Around 2.7g EPA plus DHA daily | 12 weeks |
| 2. Curcumin | Knee osteoarthritis | Positive trials, low certainty reviews | Varies hugely by formulation | 8 to 12 weeks |
| 3. Type II collagen | Knee osteoarthritis | Limited, mostly industry-funded | 40mg daily of UC-II | Up to 6 months |
| 4. Glucosamine and chondroitin | Knee osteoarthritis | Large trial found no overall benefit | 1500mg and 1200mg daily | Not recommended |
| 5. MSM | Knee osteoarthritis | Insufficient | 3g twice daily in trials | Not established |
The second column is the one people skip. Most joint supplements are marketed for arthritis as though it were one condition, and the supplement with the best evidence on this list works for the inflammatory type and does relatively little for the mechanical type that most people buying it actually have.
This is a documentary review of published trials and guidance rather than a recommendation for you personally. Around a dozen supplements marketed for joints were considered and five were kept, ranked on the quality and consistency of the evidence rather than on how well they sell.
The sources read were a 2012 meta-analysis of omega-3 in rheumatoid arthritis; a 2025 critical review of systematic reviews and meta-analyses of curcumin for knee osteoarthritis; the 2016 randomised trial of undenatured type II collagen against glucosamine plus chondroitin and placebo; the GAIT trial in the New England Journal of Medicine; a meta-analysis of MSM and DMSO in knee osteoarthritis; and NICE guideline NG226. Evidence checked in August 2026.
Three honest limits, and they matter more here than on most pages. Supplement trials are frequently funded by the manufacturer of the product being tested, which is explicitly the case for the type II collagen work. Supplements are regulated as foods rather than medicines in the UK, so what is in the bottle is not verified to the standard a drug would be. And no retail prices are quoted anywhere in this article, because prices could not be verified for this review.
Vitamin D, calcium and creatine were considered and left out, because their evidence concerns bone and muscle rather than joint pain, and are covered separately.
Check which condition it was tested in. Rheumatoid arthritis and osteoarthritis are different diseases and respond to different things. A trial in one tells you little about the other.
Check the dose against the label. This is where most supplements fail. The trial dose is frequently several times what a standard capsule provides, so the product may be the right substance at the wrong strength.
Check who paid for the trial. Manufacturer-funded trials are not automatically wrong, and independent replication is what turns a promising result into a reliable one.
Give it long enough, then actually stop. Eight to twelve weeks is a fair trial. Stopping afterwards is how you find out whether it was doing anything, since joint pain fluctuates on its own over exactly this timescale.
Ask what it is displacing. The strongest evidence in osteoarthritis is for exercise and weight management. A supplement that becomes a substitute for either is a poor trade regardless of its own evidence.
What it is
The long-chain fatty acids EPA and DHA, from oily fish or fish oil supplements, which have genuine anti-inflammatory effects through the pathways they feed into.
Who it suits
People with rheumatoid arthritis and other inflammatory arthritis. The osteoarthritis evidence is much weaker, which matters because most people buying it have osteoarthritis.
What the evidence supports
A 2012 meta-analysis found omega-3 supplementation reduced joint pain intensity, morning stiffness, the number of tender joints and the use of anti-inflammatory medication in rheumatoid arthritis. Doses above around 2.7g of combined EPA and DHA per day, sustained for more than three months, are associated with reduced NSAID use.
The dose problem
This is the practical stumbling block. A typical 1000mg fish oil capsule contains around 300mg of combined EPA and DHA, so reaching 2.7g means reading the EPA and DHA figures on the back rather than the fish oil figure on the front, and taking considerably more than one capsule.
Limits and cautions
It supports medical treatment rather than replacing it, and nobody should reduce disease-modifying medication because of it. High doses can interact with anticoagulants, so check with a pharmacist if you take one.
Why it ranks first
It has the most consistent evidence, a plausible and understood mechanism, and independent trials rather than manufacturer-funded ones.
What it is
The main active compound group in turmeric, taken as a concentrated extract rather than as the spice, which contains only a small percentage of curcuminoids.
Who it suits
People with knee osteoarthritis, which is where almost all the trials have been done.
What the evidence supports
Multiple meta-analyses report that curcumin reduces pain and WOMAC scores compared with placebo, and some find it not inferior to anti-inflammatory drugs for pain and function. Adverse events are no more common than with placebo.
Why it is not first
The certainty of the evidence is low, and a 2025 critical review of the systematic reviews themselves rated all of them as critically low quality. That is an unusual and important finding: the individual results are consistently positive and the reviews summarising them are methodologically weak, so the true effect size is genuinely uncertain.
Limits and cautions
Curcumin is poorly absorbed on its own, so formulation matters more than for most supplements, and doses are not comparable between products. It can interact with anticoagulants and may worsen reflux or gallbladder problems.
Why it ranks second
The direction of the evidence is consistently positive and its quality does not yet support a confident claim, which is exactly what second place should mean.
What it is
A small dose of collagen in its native, undenatured form, usually from chicken sternum cartilage, sold as UC-II. It is proposed to work through an immune mechanism rather than by supplying building material, which is why the dose is 40mg rather than the several grams used in collagen peptide products.
Who it suits
People with knee osteoarthritis. It is worth distinguishing from hydrolysed collagen peptides, which are a different product with a different proposed mechanism.
What the evidence supports
A multicentre randomised trial of 191 people over 180 days compared 40mg of UC-II daily against glucosamine plus chondroitin and against placebo. UC-II produced a significantly greater reduction in total WOMAC score than placebo and than the glucosamine and chondroitin combination.
Limits and cautions
The evidence base is small and much of it comes from researchers affiliated with the manufacturer, which is the reason this sits third rather than higher. Independent replication is limited. The result is genuinely promising and it is not yet established.
Why it ranks third
The trial quality is reasonable and the trial count and independence are not, which places it below curcumin despite a cleaner headline result.
What it is
The best-known joint supplements by a wide margin, both components of cartilage, taken on the reasoning that supplying them supports cartilage maintenance.
Who it suits
On current evidence and UK guidance, this is not recommended. It remains the most widely bought joint supplement in the country, which is why it appears here.
What the evidence supports
The GAIT trial randomised 1,583 people with symptomatic knee osteoarthritis to glucosamine 1500mg, chondroitin 1200mg, both, celecoxib or placebo for 24 weeks. Neither supplement alone nor in combination reduced pain significantly better than placebo overall. An exploratory subgroup with moderate to severe pain suggested the combination might help, and with only around 70 people per group the authors treated that as hypothesis-generating rather than conclusive.
What the guidance says
NICE guideline NG226 is direct: do not offer glucosamine for the management of osteoarthritis. It appears in the same recommendation as acupuncture, electrotherapy and arthroscopic lavage.
Limits and cautions
Both are well tolerated, so the argument against them is that they do not work rather than that they are harmful. Glucosamine is commonly derived from shellfish, which matters for allergy, and it can interact with warfarin.
Why it ranks fourth
It has by far the largest trial on this page, and that trial found no overall benefit, which is a clearer negative than the uncertainty attached to items two and three.
What it is
Methylsulfonylmethane, an organic sulphur compound, marketed for joint pain and often sold combined with glucosamine.
Who it suits
Nobody clearly, on current evidence. It is included because it appears in a large proportion of combination joint products.
What the evidence supports
A meta-analysis of MSM and DMSO in knee osteoarthritis found three eligible randomised trials totalling 326 patients. Two showed statistically significant but not clinically relevant pain reduction, and the meta-analysis overall confirmed a non-significant reduction in pain. An earlier systematic review concluded there was no definitive evidence to support its use.
Limits and cautions
The trials are short and small, and the reviewers called for work to establish optimum dosing, efficacy and safety, which is a fair description of a supplement whose basics remain unsettled after years of availability. It appears well tolerated at the doses studied, around 3g twice daily.
Why it ranks fifth
It has the least evidence of anything on this list, and it is frequently bought unknowingly as part of a combination product.
| Supplement | Best evidence in | Independent trials | Recommended by NICE | Main practical problem |
|---|---|---|---|---|
| Omega-3 | Rheumatoid arthritis | Yes | Not specifically addressed | Reaching the trial dose |
| Curcumin | Knee osteoarthritis | Yes | Not specifically addressed | Absorption and formulation |
| Type II collagen | Knee osteoarthritis | Limited | Not specifically addressed | Few independent trials |
| Glucosamine and chondroitin | No overall benefit found | Yes, large | No, advised against | The largest trial was negative |
| MSM | Not established | Few and small | Not specifically addressed | Very little evidence either way |
The last column is the practical summary. Two of these might work and are usually taken at the wrong dose or in the wrong formulation. One might work and has not been tested independently enough to know. One has been tested thoroughly and did not work. One has barely been tested at all.
If you have rheumatoid arthritis, omega-3 at around 2.7g of combined EPA and DHA daily is the best-supported option here, alongside your usual medication rather than instead of it.
If you have knee osteoarthritis and want to try one thing, curcumin has the most consistent positive results, with the caveat that the reviews summarising those results are weak.
If you are already taking glucosamine and feel it helps, it is safe and inexpensive, so there is no need to stop. NICE advises against recommending it, which is a different statement from telling you your experience is wrong.
If you are buying a combination product, check what is actually in it and at what dose. Combinations frequently contain small amounts of several things, none at a dose that has been tested.
If a supplement has replaced your exercise programme, that is the trade worth reversing. Nothing on this page has evidence approaching that for strengthening in osteoarthritis.
If you take anticoagulants, speak to a pharmacist before starting omega-3, curcumin or glucosamine, all of which can interact.
What has changed in clinic is the volume. People used to arrive taking one joint supplement, usually glucosamine, recommended by a relative. Now it is common to see four or five, bought after reading around the subject properly, and the reading is often better than the buying. Someone has correctly identified that omega-3 has the best evidence, and is taking one capsule a day of a product providing about a tenth of the trial dose.
The consistent gap is between the substance and the dose. Trial doses and shelf doses are different numbers for almost everything on this list, and the label is designed around the front of the bottle rather than the back. Reading EPA and DHA rather than fish oil, and curcuminoid content rather than turmeric content, changes what someone is actually taking more than switching brands ever does.
The other thing worth saying plainly is where supplements sit. For osteoarthritis the treatments with real effect sizes are strengthening and, where weight is raised, weight reduction. Supplements operate at the margin of that, and they are frequently the only thing being done. Where osteoarthritis is the diagnosis, the osteoarthritis page covers what the rest of the plan should look like.
Which joint supplement has the best evidence?
Omega-3, and specifically for rheumatoid arthritis rather than osteoarthritis. A meta-analysis found it reduced joint pain, morning stiffness, tender joint counts and anti-inflammatory medication use, at doses around 2.7g of combined EPA and DHA daily sustained for more than three months.
Does glucosamine work for arthritis?
The GAIT trial randomised 1,583 people and found neither glucosamine nor chondroitin, alone or combined, reduced knee pain significantly better than placebo overall. NICE guideline NG226 tells clinicians not to offer glucosamine for osteoarthritis. It is safe and inexpensive, so continuing one you believe helps is reasonable.
Does turmeric help joint pain?
Multiple meta-analyses report that curcumin reduces pain and WOMAC scores in knee osteoarthritis compared with placebo, with some finding it comparable to anti-inflammatory drugs. A 2025 critical review rated all the existing systematic reviews as critically low quality, so the direction is positive and the size of the effect is uncertain.
How long should I take a joint supplement before deciding if it works?
Eight to twelve weeks for most, and up to six months for type II collagen based on the trial duration. Then stop and see what happens, because joint pain fluctuates over exactly this timescale and stopping is the only way to distinguish the supplement from the natural course.
Are joint supplements safe?
The five here are generally well tolerated at the doses studied. The main cautions are interactions with anticoagulants for omega-3, curcumin and glucosamine, shellfish-derived glucosamine for people with that allergy, and curcumin in reflux or gallbladder disease. Supplements are regulated as foods rather than medicines in the UK, so content is not verified to the standard applied to drugs.
Is collagen good for joints?
Undenatured type II collagen at 40mg daily outperformed both placebo and glucosamine plus chondroitin in a 180-day randomised trial of 191 people with knee osteoarthritis. The evidence base is small and largely produced by researchers affiliated with the manufacturer, so independent replication is still needed. It is a different product from hydrolysed collagen peptides.
Should I take supplements instead of exercise for arthritis?
No. Therapeutic exercise is the core treatment for osteoarthritis in NICE guidance, and its effect is considerably larger than anything shown for any supplement on this list. Supplements work at the margin of that, and they are most often taken by people doing nothing else.
PhysioHub – Empowerment through Evidence-Based Education.
Every body is different. Book a thorough assessment and get advice that is specific to you and your goals.
Book an Appointment