The Joint Supplement Guide

The category is smaller than the shelf suggests. Strip the brand names away and most joint products are built from a list of about ten ingredients, combined in different proportions and photographed differently.

Ten families. Most products on any shelf are two or three of these plus a photograph.

The last row is where a good deal of this category lives, and it is the row worth understanding before reading any label. An ingredient can be well studied and completely unstudied in joints at the same time, because trials are run in conditions rather than in ingredients.

Alpha-lipoic acid is the clearest example. It carries two of the longest supplement trials ever run, SYDNEY 2 and the four-year NATHAN 1, both at 600 mg a day, and both are in diabetic nerve damage. Butcher’s broom is the same story in the veins: a 166-patient placebo-controlled trial measuring leg volume, and a meta-analysis of the same indication. Neither has been tested against a joint.

Two documents do most of the work here, and they are worth reading rather than quoting. One is a clinical guideline; the other is the most recent head-to-head comparison of supplements in knee osteoarthritis.

Those two documents do not contradict each other. A guideline sets a bar for what clinicians should recommend, and a network meta-analysis compares what has been measured. A supplement can show a real, modest, consistent effect in trials and still not reach the threshold at which a guideline tells clinicians to prescribe it.

For curcumin specifically the picture is the most developed in the category. A 2024 review of eleven meta-analyses found consistent reports of reduced pain and improved function. A 2025 critical appraisal is candid about how variable the underlying trials are and still finds a signal. A Bayesian network meta-analysis of curcumin dosing has gone further and modelled which amounts produce most of the effect.

The trials themselves are worth knowing by their numbers. The 2014 randomised double-blind trial used 1,500 mg a day of curcuminoids in three divided doses over six weeks. A 367-patient multicentre study set the same daily amount against ibuprofen at 1,200 mg a day for four weeks. That figure, 1,500 mg a day, is the single most useful number to carry into a shop.

The first five rows are read from the abstracts of the trials linked beside them. The last two are the conventional research amounts for ingredients this site does not sell and does not cite individual trials for.

The reason to carry those numbers is simple arithmetic. A label that names an ingredient with a printed weight beside it can be compared with the research in about ten seconds. A label that names an ingredient without a weight cannot be compared with anything at all, and that is a great deal more common than it should be.

Two practical cautions about the comparison. A percentage in front of a weight is a divisor: 10 mg of a 5 per cent extract is half a milligram of the active compound, and labels rely on nobody doing the division. And an extract is not the food it came from, so a trial run on whole seeds or whole root is not evidence for an extract of the same plant.

Supplements in this category are unusually well tolerated, which is one of the findings of the 2025 comparison: no supplement in it was associated with more adverse events than placebo. The cautions that do exist are specific, and they are about medicines somebody may already be taking.

• Turmeric and blood thinning. A dedicated interaction study examined a bioavailable curcumin formulation alongside antiplatelet agents, anticoagulants and thyroid replacement therapy. Anyone on those should say so to whoever prescribes them.
• Turmeric and the liver. A published series of ten cases links turmeric products to liver injury. A very small number against enormous use, and not zero.
• Alpha-lipoic acid and blood sugar. A 2026 review sets it out as a trigger for insulin autoimmune syndrome in susceptible people, and it applies to people who are not diabetic.
• Glucosamine and shellfish. Most glucosamine is made from shellfish, which matters to a small number of people and is rarely printed prominently.
• Stacking. Two products carrying the same ingredient is the commonest way a sensible amount becomes an unknown one, and it happens most often with turmeric.
• Surgery. Two weeks off anything studied alongside anticoagulants, and tell the surgical team what has been taken.

Beyond that, the general references are worth ten minutes before buying anything in this category. NCCIH’s guidance on using supplements wisely and the NIH Office of Dietary Supplements both cover what a supplement is regulated to do and what a label is obliged to carry. MedlinePlus on osteoarthritis is the plain-language starting point for the condition itself.

Applied to the bottle this website sells, that list produces a mixed answer and it is published rather than avoided. The lead ingredient is the right one. The grade is the research grade. The window fits a full trial. The amounts are not printed, which means the third step cannot be carried out at all. The scorecard turns that into eight checks with a pass or a fail against each, and applies them to Artivorin at the end. It fails three of them outright and cannot answer a fourth.