Food supplements for the eyes: what works, what does not, and for whom

Supplements “for the eyes” sell very well, and the essence of their promise rests on a misunderstanding. The formula known as AREDS2 was assessed in a large randomised trial — but in patients who already had age-related macular degeneration, and to slow its progression, not to stop it appearing. This article sorts out what has been demonstrated, what has not, and what can do harm.

UNDERSTANDING

A whole pharmacy shelf, and very little evidence

In a French pharmacy, the shelf space given over to vision now takes up as much room as the one given over to sleep. The boxes all look alike: a stylised eye, molecules with reassuring names, a deliberately vague promise of “support for visual function”. The patients who bring them to me almost always ask the right question: is this any use, for me?

The answer is uncomfortable for everyone: it depends entirely on what you already have. The same tablet may have a demonstrated benefit in one person, no demonstrated benefit in her neighbour, and a real drawback in her brother who smokes. A supplement is not a neutral food added as a precaution: it is a pharmacological dose of micronutrients, taken every day for years.

An ophthalmologist’s job is neither to sell these products nor to mock the people who buy them, but to say what the trials measured and in whom.

AREDS and AREDS2: what exactly are we talking about

Two trials, and a formula that became a reference

The original formula, from the Age-Related Eye Disease Study (AREDS), combines antioxidants and zinc: vitamin C 500 mg, vitamin E 400 IU, beta-carotene 15 mg, zinc 80 mg, with copper. AREDS2 tested two additions: lutein 10 mg with zeaxanthin 2 mg, and long-chain omega-3, 350 mg of DHA and 650 mg of EPA.

AREDS2 is a serious trial, and that must be said: multicentre, randomised, double-blind, placebo-controlled, with a 2 × 2 factorial design, run from 2006 to 2012 in 4,203 participants aged 50 to 85. Methodological quality is not the problem: the problem is how it is read.

In whom, precisely, the formula was assessed

The participants were not recruited from the general population. Large drusen in both eyes were required, or large drusen in one eye and already advanced AMD in the other: 66% had bilateral drusen, 34% had one eye already affected by an advanced form. Nobody in this trial had a normal macula.

The primary endpoint was the occurrence of advanced AMD — the neovascular form or central geographic atrophy — with a median follow-up of five years. The trial measured neither visual comfort, nor fatigue in front of a screen, nor the protection of a healthy eye.

What the trial showed, and what it did not show

Let us begin with the result that is quoted least, because it is negative. At five years, the probability of progression to advanced AMD was 31% on placebo, 29% on lutein and zeaxanthin, 31% on DHA and EPA, 30% on the combination. The hazard ratio for lutein and zeaxanthin was 0.90 (98.7% confidence interval: 0.76 to 1.07), which is not significant. Adding these molecules to the AREDS formula did not reduce the risk any further.

What does remain established is the effect of the original antioxidant formula. The most recent Cochrane review, which brings together 26 trials and 11,952 people who have AMD, finds a reduction in progression to advanced AMD: an odds ratio of 0.72 (0.58 to 0.90), moderate certainty of evidence.

But this review does what the advertisements never do: it translates that relative figure into absolute terms, stage by stage. In people with intermediate AMD, it represents about 78 cases of progression avoided per 1,000 people treated (26 to 126). In those with early AMD, it falls to about 4 cases avoided per 1,000 (1 to 6). Same tablet, same relative percentage, an altogether different practical value.

An analysis carried out after the event on the AREDS and AREDS2 data further suggests that supplementation slows the progression of geographic atrophy towards the centre of the macula: 50.7 micrometres per year against 72.9. This is relevant, since it is the centre that carries fine vision, but a post hoc analysis is a hypothesis generated after the fact.

“And if I have nothing?” — the question that decides everything

This is where it all hangs, and the answer is direct: in a person who does not have AMD, the AREDS2 formula has no demonstrated benefit. None of these trials included participants with a normal macula in order to test prevention. The question was not put for this formula, so it has not been settled.

This reading is not mine alone. The title of the Cochrane systematic review devoted to antioxidant and mineral supplements is about slowing progression, not about prevention (PMID 37702300). A review devoted to the controversies around supplements in ophthalmology reaches the same conclusion (PMID 26350529): good-quality data support an AREDS-type formula in patients at moderate to high risk of advanced AMD, and they are insufficient to recommend the everyday use of commercial products outside that situation.

I am told that an absence of evidence is not evidence of absence, and that is correct. But we are dealing with a product taken daily, for life, with a real cost and adverse effects that are not nil: the burden of proof then lies with the product, not with the patient.

If AMD has already been diagnosed, the discussion changes in nature: the stage, the state of the fellow eye and the rate of progression become decisive, and they are assessed in consultation, with imaging to hand. That is the subject of the follow-up of dry AMD.

Beta-carotene: the proof that a supplement can do harm

This is the most important chapter of this article, and the one the shelves ignore most thoroughly. In AREDS2, a second randomisation compared the formula with and without beta-carotene. The investigators recorded 23 lung cancers (2.0%) in the group receiving beta-carotene, against 11 (0.9%) without, occurring for the most part in former smokers. It was this observation that led to beta-carotene being taken out of the formula.

This signal is not an isolated one, and that is what gives it weight. A meta-analysis of eight randomised trials covering 167,141 participants finds a relative risk of lung cancer of 1.16 (1.06 to 1.26) with supplementation, and of 1.21 (1.08 to 1.35) in smokers and in workers exposed to asbestos. A more recent meta-analysis finds 1.19 (1.08 to 1.32), that is an increase of about 19% in relative terms. A third also reported a higher risk of stomach cancer at doses of 20 to 30 mg a day.

One distinction is indispensable, and it is a liberating one: the beta-carotene in carrots, apricots and spinach is not concerned. Recent work clearly separates dietary intake, which is associated with a reduction in cancer risk, from high-dose supplementation in the heavy smoker. It is not the molecule that is at fault: it is the dose, the form and the person’s profile.

The replacement was made in favour of lutein and zeaxanthin, and the secondary analyses checked that the substitution cost nothing in effectiveness: in direct comparison, they did at least as well as beta-carotene on progression to late AMD (hazard ratio 0.82, from 0.69 to 0.96). Removing a problematic molecule without losing the benefit: this is a good example of medicine correcting itself.

The lesson goes beyond ophthalmology. “Natural” does not mean harmless, and “available without prescription” does not mean without consequence: high-dose vitamin E has been associated with an increase in prostate cancer (PMID 21990298). A micronutrient is only useful where it is lacking.

Lutein and zeaxanthin: a real biological effect, a limited reach

Lutein and zeaxanthin are carotenoids of the xanthophyll family. The body cannot make them: they come from food, above all from green leafy vegetables and egg yolk. They concentrate at the centre of the retina, where they form the macular pigment, which filters blue light and plays a local antioxidant role. The hypothesis is an attractive one, hence its commercial success.

And it works, at least on the marker. A meta-analysis of 20 randomised trials covering 938 patients with AMD and 826 healthy subjects shows an increase in macular pigment optical density: 0.07 (0.03 to 0.11) in the patients, 0.09 (0.05 to 0.14) in the healthy subjects, with a dose-effect relationship.

That leaves the question the patient cares about: does it change his vision or the course of his disease? The answer is far cooler. Against a control, lutein combined with zeaxanthin goes with a risk of progression to late AMD of 0.94 (0.87 to 1.01) and a risk of losing 15 letters or more of 0.98 (0.91 to 1.05), at low certainty of evidence. On visual function, a meta-analysis of eight trials in 1,176 patients finds a gain of 0.04 logMAR, that is about two letters.

Two letters is not nothing for a researcher and not a great deal for someone who wants to read his newspaper again. Both statements have to be held together: raising a biomarker is demonstrated, transforming a patient’s life is not. Confusing the two is the commonest manoeuvre in the marketing of supplements.

Omega-3 and dry eye: trials that disappoint

The reasoning is the same: long-chain omega-3 have anti-inflammatory properties and form part of the lipid film secreted by the meibomian glands, whose dysfunction is the leading cause of dry eye. On paper, it ought to work. In patients, it is far less clear-cut.

The Cochrane review devoted to the subject brought together 34 randomised trials and 4,314 adults from 13 countries, with follow-up ranging from one to twelve months. On symptoms, it concludes that the effect is small or nil against placebo: a mean difference of −2.47 points, interval from −5.14 to 0.19, therefore compatible with no effect at all, at low certainty of evidence. On the Schirmer test, the increase is probable but the authors did not judge it clinically significant. More than half of the trials were at high risk of bias.

A more recent meta-analysis, on 19 trials and 4,246 patients, concludes on the contrary that there is a significant improvement in symptoms, in tear film break-up time and in osmolarity. This contradiction is worth looking at squarely: it comes from the heterogeneity of populations, doses and durations, and from a problem with the comparator, several trials having used olive oil as their placebo — and olive oil is not inert.

A more convincing line is emerging: the benefit would be concentrated in people who are genuinely deficient. In a trial covering 470 heavy screen users given supplements against 480 on placebo for six months, 81% had a low omega-3 index at the outset, and the improvement there was distinctly more marked. Correcting a deficiency, then, rather than supplementing everybody.

In practice, I do not discourage a patient who wants to try, but I refuse to let the capsule take the place of what has a better level of evidence: treating meibomian gland dysfunction, taking care of eyelid hygiene, correcting the working environment and the blink rate. These measures are set out in our article on dry eye linked to home working and screens.

On the retinal side, the verdict is more clear-cut: in AREDS2, adding DHA and EPA had no effect at all on the progression of AMD (hazard ratio 0.97, from 0.82 to 1.16). Observational work nevertheless points the other way: in a cohort of 258,350 people free of AMD followed for nearly thirteen years, a higher plasma level of omega-3 went with a lower risk of diagnosis, a result supported by a Mendelian randomisation analysis. That justifies new trials, not a recommendation.

Vitamins and cataract: the gap between observation and trial

Cataract offers the most instructive illustration of the gap between two kinds of evidence. A meta-analysis bringing together 12 cohorts and 8 randomised trials finds, in the cohorts, distinctly favourable associations: 0.80 for vitamin C, 0.81 for vitamin A, 0.81 for lutein or zeaxanthin, 0.90 for vitamin E.

Except that the same meta-analysis then examines the randomised trials, and the result evaporates: supplementation with vitamin E gives 0.97 (0.91 to 1.03) and with beta-carotene 0.99 (0.92 to 1.07), which is to say nothing at all. A dedicated trial in 11,267 men followed for an average of 5.6 years, with 389 documented cataracts, reaches the same conclusion: 1.02 (0.84 to 1.25) for vitamin E and 0.91 (0.75 to 1.11) for selenium.

The most telling detail comes from a meta-analysis devoted to vitamin E alone: dietary intake is associated there with a relative risk of 0.73, intake by supplementation alone with 0.92 (0.78 to 1.07), which is not significant. Same molecule, same disease, two opposite results depending on whether it comes from a plate or from a capsule.

The explanation comes down to one word: confounding. Heavy consumers of fruit and vegetables differ from the rest in a host of characteristics — physical activity, smoking, income, access to care. Observation picks up that whole bundle, the randomised trial isolates the molecule; when the two diverge, it is the trial that settles the matter. What seems to count is the diet as a whole, the subject of our article on diet and the prevention of cataract.

So, for whom? A reading situation by situation

Here is how I sum the question up in consultation, according to the actual eye situation. None of these lines replaces an examination: they are there to prepare the conversation.

• Intermediate AMD, or an advanced form in one eye. This is the trial population, and it is here that the benefit is best documented. The formula without beta-carotene is worth discussing openly with your ophthalmologist.
• Early AMD, a few isolated drusen. The absolute benefit to be expected is very small, in the order of a few cases avoided per thousand people treated.
• No macular abnormality. No demonstrated benefit. A family history increases your risk of AMD, but it has never been shown that a supplement changes anything about it.
• Smoker or former smoker. Check that there is no beta-carotene or provitamin A in the composition, whatever the reason for taking it.
• Dry eye. Omega-3 can be tried, without expecting any transformation, and never in place of treating the cause.
• Early cataract. No supplement has been shown to slow cataract; the only effective treatment remains surgical, when the time comes.

A few points of caution should be added. The adverse effects most often reported in the trials are digestive. High-dose zinc has been associated with more hospital admissions for genito-urinary problems (PMID 23152201). Finally, stacking products with similar compositions exposes people to cumulative doses that nobody is keeping track of: bring the boxes to the consultation.

Cachan practice · Tel. +33 1 45 47 08 11

What it is reasonable to do, with or without a supplement

Concluding with a “none of it is any use” would be dishonest. What the work on nutrition and AMD shows is that the soundest levers are not in a box: the combination of several behaviours — not smoking, keeping active, adopting a Mediterranean-style diet — is associated with a lower frequency of AMD.

• Stopping smoking is by far the most effective measure there is on the risk of AMD.
• A diet rich in green vegetables, brightly coloured fruit and oily fish supplies the same molecules, at physiological doses and with no risk of overdose.
• Sunglasses that filter ultraviolet light are a simple and inexpensive measure.
• Controlling blood pressure and weight acts on the retinal microcirculation.
• A regular fundus examination after the age of 55 picks up drusen at the stage where the question of a supplement finally becomes relevant.

If you leave here without buying anything, this article will have done its job. The right question has never been “which supplement should I take”, but “am I in the situation where something has been demonstrated”. For the great majority of readers, the answer is no.

Frequently asked questions

Cachan practice · Tel. +33 1 45 47 08 11

Scientific sources

• Age-Related Eye Disease Study 2 Research Group. Lutein + zeaxanthin and omega-3 fatty acids for age-related macular degeneration: the AREDS2 randomized clinical trial. JAMA. 2013;309(19):2005-15. PMID 23644932
• AREDS2 Research Group. The Age-Related Eye Disease Study 2 (AREDS2): study design and baseline characteristics (AREDS2 report number 1). Ophthalmology. 2012. PMID 22840421
• Chew EY, et al. Secondary analyses of the effects of lutein/zeaxanthin on age-related macular degeneration progression: AREDS2 report No. 3. JAMA Ophthalmol. 2014;132(2):142-9. PMID 24310343
• Evans JR, Lawrenson JG. Antioxidant vitamin and mineral supplements for slowing the progression of age-related macular degeneration. Cochrane Database Syst Rev. 2023;9:CD000254. PMID 37702300
• Keenan TDL, et al. Oral antioxidant and lutein/zeaxanthin supplements slow geographic atrophy progression to the fovea in age-related macular degeneration. Ophthalmology. 2025;132(1):14-29. PMID 39025435
• Lawrenson JG, Grzybowski A. Controversies in the use of nutritional supplements in ophthalmology. Curr Pharm Des. 2015;21(32):4667-72. PMID 26350529
• Carneiro Â, Andrade JP. Nutritional and lifestyle interventions for age-related macular degeneration: a review. Oxid Med Cell Longev. 2017;2017:6469138. PMID 28154734
• Ma L, et al. Lutein, zeaxanthin and meso-zeaxanthin supplementation associated with macular pigment optical density. Nutrients. 2016;8(7):426. PMID 27420092
• Liu R, et al. Lutein and zeaxanthin supplementation and association with visual function in age-related macular degeneration. Invest Ophthalmol Vis Sci. 2014;56(1):252-8. PMID 25515572
• Downie LE, et al. Omega-3 and omega-6 polyunsaturated fatty acids for dry eye disease. Cochrane Database Syst Rev. 2019;12:CD011016. PMID 31847055
• Wang WX, Ko ML. Efficacy of omega-3 intake in managing dry eye disease: a systematic review and meta-analysis of randomized controlled trials. J Clin Med. 2023;12(22):7026. PMID 38002640
• Bhargava R, et al. Omega-3 fatty acids supplements for dry eye – are they effective or ineffective? Indian J Ophthalmol. 2023;71(4):1619-25. PMID 37026312
• Souied EH, et al. Omega-3 fatty acids and age-related macular degeneration. Ophthalmic Res. 2015;55(2):62-9. PMID 26610051
• Xue CC, et al. Omega-3 fatty acids as protective factors for age-related macular degeneration: prospective cohort and Mendelian randomization analyses. Ophthalmology. 2025;132(5):598-609. PMID 39662686
• Jiang H, et al. Dietary vitamin and carotenoid intake and risk of age-related cataract. Am J Clin Nutr. 2019;109(1):43-54. PMID 30624584
• Christen WG, et al. Age-related cataract in men in the selenium and vitamin E cancer prevention trial eye endpoints study: a randomized clinical trial. JAMA Ophthalmol. 2015;133(1):17-24. PMID 25232809
• Zhang Y, et al. Vitamin E and risk of age-related cataract: a meta-analysis. Public Health Nutr. 2015;18(15):2804-14. PMID 25591715
• Wei L, et al. Association of vitamin C with the risk of age-related cataract: a meta-analysis. Acta Ophthalmol. 2016;94(3):e170-6. PMID 25735187
• Kordiak J, et al. Role of beta-carotene in lung cancer primary chemoprevention: a systematic review with meta-analysis and meta-regression. Nutrients. 2022;14(7):1361. PMID 35405977
• Zhang Y, et al. Association between beta-carotene supplementation and risk of cancer: a meta-analysis of randomized controlled trials. Nutr Rev. 2023;81(9):1118-30. PMID 36715090
• Druesne-Pecollo N, et al. Beta-carotene supplementation and cancer risk: a systematic review and meta-analysis of randomized controlled trials. Int J Cancer. 2010;127(1):172-84. PMID 19876916
• Bates CA, et al. Investigating the relationship between beta-carotene intake from diet and supplements, smoking, and lung cancer risk. Food Chem Toxicol. 2024;194:115104. PMID 39522798
• Harvie M. Nutritional supplements and cancer: potential benefits and proven harms. Am Soc Clin Oncol Educ Book. 2014:e478-86. PMID 24857143
• Christen WG, et al. Effects of multivitamin supplement on cataract and age-related macular degeneration in a randomized trial of male physicians. Ophthalmology. 2014;121(2):525-34. PMID 24268861
• Klein EA, et al. Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA. 2011;306(14):1549-56. PMID 21990298
• Evans JR, Lawrenson JG. Antioxidant vitamin and mineral supplements for slowing the progression of age-related macular degeneration. Cochrane Database Syst Rev. 2012;11:CD000254. PMID 23152201 (2012 edition, cited for the genito-urinary signal with zinc)

This article is intended for information and does not replace a medical consultation. Indications, expected benefits and risks depend on each individual situation. For any question about your vision, consult an ophthalmologist.