Vision Supplement Guide: What The Evidence Shows

The category exists because of one finding in one population, and almost everything sold in it since has borrowed that finding's reputation.

The population was people with intermediate age-related macular degeneration, or with advanced disease in one eye. The finding was that a specific high-dose formula reduced progression to advanced disease over roughly six years. That is a narrow, real and valuable result, and it is not a general statement about eyes.

Everything else sold in this category falls into a second group with a much weaker evidence base: capsules for everyday comfort, for screen hours, for tired eyes at the end of a working day. Those are legitimate things to want and they are not what the famous trials measured.

Knowing which group a product belongs to is the single most useful thing a shopper can establish, and the label usually says. A formula built on the trial amounts is unmistakable: high-dose vitamin C and E, zinc, copper and two carotenoids. Anything else is a multivitamin with an eye framing, which may still be worth buying for different reasons.

Amounts read from the published trial reports, each of which is cited below.

The original report is the document the whole category stands on, and the ten-year follow-up is what makes it durable rather than a single positive result. Both are worth opening.

The negatives matter as much. The cataract arm of the same trial found nothing at all, which means the formula's reputation belongs to one condition rather than to eyes in general. And the Cochrane review on prevention concluded that antioxidant supplements do not stop macular degeneration developing in people who do not have it.

Put together, the honest summary is narrow and useful: a specific high-dose formula slows progression in people who already have intermediate disease, does not prevent the disease starting, and does nothing for cataract. The 2023 Cochrane review on slowing progression is the current synthesis, and a network meta-analysis ranks the individual nutrients for anyone who wants to know which parts carry the weight.

These two carotenoids are not antioxidants in general; they are the specific pigments that accumulate in the macula and give it its colour. That anatomical fact is why they have the best evidence in this category and why every serious modern formula carries them.

The AREDS2 secondary analysis found them a better choice than beta-carotene, and the ten-year report confirmed the substitution. The amounts were 10 mg and 2 mg a day, and those two figures are the benchmark any label should be read against.

The case outside macular disease is thinner but it is not nothing, and it is the part relevant to most shoppers. A one-year contrast-sensitivity trial, a trial in daily screen users and a study in people with long screen hours all report changes in macular pigment or in visual performance measures. A meta-analysis of xanthophyll intake pools the wider picture.

For a shopper the practical rule is short. A formula sold for screen comfort without either carotenoid is missing the two ingredients with the most direct evidence in that setting, whatever else it carries. Dark leafy greens, sweetcorn and egg yolk are the dietary sources, and they are cheaper than any capsule.

Fish oil has a strong reputation in eye supplements and a weak record in eye trials, and the gap is worth knowing about before paying for it.

AREDS2 tested 350 mg of DHA plus 650 mg of EPA daily on top of the original formula and found no additional benefit for macular degeneration. That is a well-powered negative result in exactly the population where a benefit was expected.

The DREAM trial went at the other common claim, giving 3,000 mg a day of omega-3 for dry eye disease over twelve months and finding no significant difference against an olive oil comparator.

Neither trial says fish oil is useless; both say it did not do what an eye supplement sells it for, at generous amounts, in well-run studies. A shopper who wants omega-3 for cardiovascular reasons has a separate argument and should read a separate literature.

Zinc is the ingredient in the AREDS formula with the most independent support, and it is also the one that created a second problem the trial had to solve.

At 80 mg a day of zinc oxide, sustained over years, zinc competes with copper for absorption and depresses copper status. The consequences are documented: a case of reversible blood counts and a report of marrow suppression are the kind of thing that follows prolonged high intake. AREDS included 2 mg of cupric oxide specifically to prevent it.

The Institute of Medicine sets the adult tolerable upper intake at 40 mg a day, which is half the AREDS amount, and that tension is one reason AREDS2 tested a 25 mg arm.

The rule this gives a shopper is precise and easy to apply. If a label carries a high zinc amount, look for copper; if it does not, the formula is either following the trial carelessly or is at a low enough zinc amount that the question does not arise. Both answers are informative, and only one of them is a problem.

Folate, vitamin B6 and vitamin B12 lower homocysteine, and raised homocysteine is associated with small-vessel disease, including in the retina. That is the reasoning behind putting the trio in an eye formula.

Unusually for this category, it was actually randomised. The Women's Antioxidant and Folic Acid Cardiovascular Study followed 5,205 women for an average of 7.3 years on folic acid 2.5 mg, pyridoxine hydrochloride 50 mg and cyanocobalamin 1 mg daily, and reported fewer cases of macular degeneration in the treated group.

A 2026 review of the B vitamins in macular degeneration sets out where that leaves things: one trial, an endpoint that was not the primary one, and no replication in the two decades since. It is a genuine signal rather than a proven effect.

Two practical notes for a shopper. Those three amounts are far above what a multivitamin carries, so a label with folate, B6 and B12 on it is not thereby a WAFACS formula. And high folic acid intake has its own consideration: a review of the interaction explains why folate can correct the blood picture of B12 deficiency while the nerve problem continues, which is why the two belong together.

This is the one safety question in the category that produced real harm in real trials, and it is also the one most often repeated without its amounts attached.

The CARET trial gave 30 mg of beta-carotene a day with retinyl palmitate to smokers and asbestos-exposed workers and was stopped early after more lung cancers appeared in the treated group. The ATBC trial found the same direction at 20 mg a day. A later meta-analysis confirmed that the effect belongs to high-dose supplementation in that population.

AREDS2 removed beta-carotene from the formula for this reason, replacing it with the two carotenoids. That is the single clearest example in this field of a trial changing a product because of a safety finding.

What the warning does not say is that any amount of beta-carotene is dangerous to anybody. It is a dose-dependent finding in smokers at twenty to thirty milligrams a day. A multivitamin carrying a fraction of a Daily Value is a different exposure, and a reader deciding whether a warning applies to a label needs the number rather than the headline.

Digital eye strain is real and well characterised: tired eyes, dryness, intermittent blur and headache after sustained near work. A 2024 review and the TFOS Lifestyle report describe the same mechanisms, and they are mechanical: a reduced blink rate, sustained accommodation and an unstable tear film.

That shapes what helps. A meta-analysis of every tested intervention is the best single source, and the measures with support are the ones aimed at those mechanisms: blink breaks, screen distance and height, ambient lighting, and an up-to-date correction for the working distance.

Two popular answers do less than they are sold as doing. The Cochrane review of blue-light filtering lenses found no benefit for eye strain, and a paper examining the 20-20-20 rule points out that its three numbers were never derived from a trial, which does not make the habit useless but does make the precision spurious.

The nutritional argument in this setting is the carotenoid one and nothing else. That is why a formula sold for screen hours should be read first for lutein and zeaxanthin, and why an optometrist who checks your working-distance correction is likely to change more than any capsule will.

Every product in this category uses the same grammar, and understanding it saves a shopper a great deal of time.

A dietary supplement may describe an effect on a structure or a function of the body. It may not say it treats, cures, mitigates or prevents a disease, because a product that says so has become an unapproved drug rather than a supplement. That single rule generates the entire vocabulary of the shelf: supports, promotes, helps maintain, contributes to.

So the language tells you almost nothing about the evidence. A formula built on the trial amounts and a multivitamin with an eye picture on it are obliged to use the same verbs, and the strongest product on a shelf cannot say anything stronger than the weakest one.

What does differentiate them is on the back of the pack. The amounts, the presence or absence of the two carotenoids, whether a blend hides its split, and whether the seals name anybody. An analysis of FDA warning letters is a reminder of what happens at the far end of the category, and the products involved are almost always the ones that publish least.

That is what the second guide on this site turns into seven questions you can ask without remembering any of this. The scorecard is here.