Doctor, What Vitamins Should I Take for My Eyes?

Today, more than ever, ophthalmologists are being asked to answer patients'questions about vitamin supplements and nutrition. The importance of dietand eye health is not new. The need to obtain adequate vitamin A to preventxerophthalmia and nightblindness, particularly where malnutrition is rampant,has been known for decades (reviewed by Underwood and Arthur¹).More recently, interest has been directed at whether nutritional supplementsmight prevent loss of vision caused by degenerative conditions that becomemore common as we age, such as cataract and macular degeneration.

The benefits of nutritional supplements may have broad public healthimportance. Results of the Age-Related Eye Disease Study (AREDS)² suggestthat nutritional supplements may be one of the most promising means discovered,to date, of delaying end-stage age-related macular degeneration (AMD), themost common cause of blindness among older people in developed countries.³ There is some interest in the possibility that supplementsmight slow the progression of diabetic retinopathy,⁴ thenumber 1 cause of blindness among working-age people.³ Moreover,some studies (reviewed herein) suggest that supplements might slow the developmentof cataract, which affects more than half of us by age 75 years.^5,6 Inaddition, surgery for cataracts is expensive, accounting for more than 12%of the Medicare budget, which was last evaluated in 1992.⁷ Asthese age-related conditions become more prevalent in a population that isaging,⁸ the potential public health benefitsof supplements are large. However, scientific evidence to support these benefitsis stronger in some cases than others.

The purpose of this review is to provide guidelines for clinical practiceon recommending the use of nutritional supplements for reducing the developmentof eye diseases that are common among older people, based on the current evidence.While supplementation may be considered in treatment of more rare inheritedretinal degenerations, this will not be considered in the current review.Benefits of several specific types of supplements that are commonly availablefor slowing the development of common eye diseases and risks are consideredin separate sections that follow. We also discuss issues for clinicians toconsider as more scientific evidence becomes available, from the many studiesthat are expected to emerge during the next few years.

Ultimately, any physician hopes to improve the overall health and well-beingof patients, rather than merely focus on eye health. Therefore, we will alsobriefly consider the evidence that describes the benefits and risks of supplementsto overall health. For a thoughtful and more thorough discussion of this largertopic, we refer readers to the recent article by Willett and Stampfer.⁹

Extreme deficiencies of many vitamins and minerals have been shown tocause cataract or retinal dysfunction in experimental animals, particularlyunder extreme experimental conditions in the laboratory (previously reviewed^10,11). In humans, the results of epidemiologicstudies are needed to determine whether more modest fluctuations of vitaminsand minerals influence age-related degenerative conditions under the conditionsthat are unique to people. Studies of large and diverse populations providean estimate of the magnitude of impact that changes in vitamins and supplementshave on the risk of common eye diseases, relative to other medical or lifestylechanges that can be recommended to patients (such as stopping smoking, controllinghypertension, losing weight, etc). Results of observational investigationsand clinical trials are summarized in the following sections.

Multivitamins usually contain all essential vitamins that one wouldexpect to get from foods at levels that meet the nutrient requirement of nearlyall healthy individuals. These nutrient levels are the highest requirementof the recommended dietary allowances (RDAs) for specific age and sex groups.¹² Many, but not all, also contain the essential mineralsat levels that we typically get from food. The RDAs are set at levels thatare judged by panels of scientists to be those needed to promote health inmost groups of healthy people. However, they may not be adequate for individualswith unique needs caused by the presence of disease.

Observational studies in 8 different populations indicate lower ratesof cataract or cataract extraction among people who use multivitamin supplements,compared with only 2 studies that observed no association (previously discussed¹³ and more recently reported¹⁴).However, in the absence of randomized controlled clinical trials, there isno proof, at this time, that multivitamins lower the risk of cataract. Thepossibility exists that other aspects of a healthy lifestyle among supplementusers explain the lower rates of cataracts in the observational studies.

Unfortunately, we may never have the guidance of clinical trials toanswer the questions of benefits of multiple supplements on cataract in theAmerican population. If multivitamins do lower risk, as the observationalstudies consistently suggest, then many years may be required to observe abenefit in clinical trials—yet, such trials are generally conductedfor less than 10 years.

To date, there are results of only one clinical trial of multivitaminsupplements. This trial, in a malnourished population in China,¹⁵ testedthe influence of using multivitamins for 5 years on the prevalence of cataract.A lower prevalence of cataract was observed in users than nonusers of multivitamins,among persons 65 to 74 years of age (but not among persons 45 to 64 yearsof age). The results of this 5-year study suggest that short-term effectsmay be possible. Considered in conjunction with observational studies in whichonly long-term supplementation was associated with lower risk, these resultssuggest that short-term effects might be limited to malnourished populations.This result will not likely be duplicated in the United States, because peoplewith poor diets are not usually subjects in clinical trials. If the influenceof multivitamins on cataract is gradual, over many years, or limited to peoplewith very poor diets, then the possible benefits of multivitamins on cataractare not likely to be experimentally proved.

A 9-year, randomized, placebo-controlled clinical trial of the influenceof multivitamin supplementation on cataract development and progression iscurrently under way in Italy in about 1000 people.¹⁶ Thisstudy is likely to provide new insights. Also, in the next 5 to 10 years,the results from many long-term prospective observational studies that arecurrently under way will provide future insights about whether multivitaminsthemselves or other associated lifestyles are responsible for lower cataractrisk among people who use supplements. If people who take multivitamin supplementsconsistently have a lower risk of cataract across many different segmentsof a population, then there is a lower possibility that the apparent beneficialeffect is caused by other unrevealed factors.

In contrast, there is no evidence that use of multivitamins slows theonset or progression of AMD, although it has been investigated in severalpopulations (previously discussed¹⁷ and morerecently reported¹⁸). This may be because peoplewho begin to show signs of developing macular degeneration may decide or beadvised to start taking nutritional supplements. This change in supplementuse (depending on having signs of the conditions) may confound relationshipsof long-term supplement use to age-related maculopathy. This would make itdifficult to observe a protective association if one existed. Such uncontrolledconfounding or noise, inherent in any epidemiologic study, may overwhelm anysmall beneficial effect. It remains conceivable that multivitamins may ensureadequacy of intake of several nutrients that are important to the health ofthe retina and retinal pigment epithelium. However, evidence of beneficialeffect may be hard to generate because supplementation is common, particularlyin people who have a family history of early signs of macular degeneration.

Considering the evidence that multivitamins might help delay cataract,even if there is a lack of evidence that multivitamins influence macular degeneration,why not recommend a general multivitamin for health "just in case"? The ideathat multivitamins are good for general health remains controversial. Thereis currently some evidence that suggests multivitamins result in fewer daysof illness due to infection.¹⁹ However, evidencein observational studies that multivitamin users have lower rates of commonchronic diseases such as coronary disease²⁰ and/orcolon cancer²¹ is scarce. Evidence of effectivenessof multivitamins in clinical trials is lacking. Fletcher and Fairfield²² recently argued that subclinical deficiencies ofsome nutrients such as folate, vitamin D, and vitamin B are common in olderpeople and may increase risk of cardiovascular disease, cancer, and osteoporosis;they recommend the use of multivitamins despite the lack of strong evidenceof effectiveness.

There is no direct evidence to support the notion that multivitamin-mineralsupplements (those that meet 100% of RDA) pose a health risk. However, specialcautions may apply in certain circumstances; for example, there may be a riskof people getting high levels of vitamin A in the diet (found in meat, liver,and milk sources) or in breakfast cereals or meal replacement beverages. Evidencesuggests that excessive dietary intake of preformed vitamin A (retinol) isassociated with increase risk of osteoporosis and hip fractures.^23,24 Thesafe upper limit for intake of vitamin A is 3000 µg retinal equivalents.²⁵ This is roughly 4 times the RDA of 700 µg/dfor women and 800 µg/d for men.²⁵ Multivitaminpreparations of vitamin A typically provide 800 µg/d, well below thesafe upper limit. The average intake of vitamin A from foods in a nationallyrepresentative population in the last decade was about 1000 µg/d. Thus,for average Americans, taking a multivitamin supplement still results in alevel of intake only 2 times the RDA. However, many people have levels intheir diets that are higher. For example, among women in the Nurses' HealthStudy, the average vitamin A (retinol) intake from foods in women in the upper20 percentiles for intake of vitamin A was 2500 µg/d.²⁴ Usinga multivitamin that contains the RDA for vitamin A along with a diet richin vitamin A may result in a total level of intake above the safe upper limit.

A final concern is giving patients the impression that multivitaminswill provide adequate micronutrients, reducing the effort to obtain them infoods. Micronutrients in foods may be more bioavailable. This may be the casefor zinc, which is better absorbed from animal sources.²⁶

An advantage of consuming the nutrients we need from foods instead ofsupplements is that food provides different chemical forms of a nutrient thatmay be important. For example, supplements contain only the vitamin E compound α-tocopherol,which is thought to have the highest vitamin E activity, even though γ-tocopherolconstitutes the majority of vitamin E in the diet. There is mounting evidencefor unique or more potent properties of γ- and δ-tocopherols ina variety of mechanisms that may be involved in preventing the pathogenesisof chronic diseases (reviewed previously²⁷).Moreover, supplementation with α-tocopherol reduces serum and tissuelevels (previously reviewed^27,28)of γ- and δ-tocopherols that we get from foods. Therefore, thenutrient form found in supplements may not be sufficient and may actuallyreduce the bioavailability of other nutrients, although there is currentlylittle evidence of specific health risks.

A third advantage to relying on the foods we eat, rather than supplements,is that foods contain other chemicals that have health benefits. Currently,evidence is building to support the benefits of many food components not containedin supplements. For example, the plant pigments lutein and zeaxanthin maybe important for eye health and general health (discussed in a later section).There are likely to be many other important food chemicals of which we arecurrently unaware. Thus, multivitamins may give patients a false sense ofsecurity about their nutritional status that will reduce the intake of potentiallyimportant food components.

There is substantial evidence that the potential for oxidative stressin the retina and lens is high and that oxidative stress contributes to thedevelopment of cataract and degenerative changes in the retina. Many supplementsare marketed to prevent age-related eye diseases that contain antioxidantssuch as vitamin C (given as ascorbic acid) and vitamin E, carotenoids, andzinc at high levels. Levels of other nutrients in these supplements are generally2 to 10 times the RDA. Selenium, which, at high doses, has been shown to causecataract,²⁹ is contained in lower levels.

There are some data to support a beneficial effect of high-dose antioxidantsupplements in delaying or preventing cataracts, but overall the body of evidenceis weak. In 2 large clinical trials,^30,31 high-doseantioxidants did not delay cataract over approximately 6 years. However, ina smaller trial (158 persons who completed the 3-year study), antioxidantsupplements did delay early lens opacification.³² Inthe latter trial, blood levels of antioxidants were higher in the supplementedgroup, and lower in the control groups, than in the AREDS,³⁰ whichmight explain the differing results. However, confirmation of this smallertrial is important to rule out the possibility that chance or some featuresof the study design, implementation, or population explain the positive resultsthat differ with the results of other clinical trials. Also, while this trialand results of several observational studies have reported lower occurrenceof lens opacities among people taking antioxidant supplements, they did notdemonstrate benefits beyond those that would be possible from general multivitamins.

In contrast to cataract, there is evidence from a large, randomizedclinical trial that antioxidant supplements may benefit people with AMD.² In the AREDS, patients taking a specific high-doseantioxidant supplement containing ascorbic acid (500 mg), vitamin E (400 IU),and beta carotene (15 mg) along with high-dose zinc (80 mg as zinc oxide togetherwith 2 mg of copper as cupric oxide)²⁹ hada 28% lower rate of clinical progression to end-stage AMD (either neovascularmacular degenerative or geographic atrophy) if they already had intermediateor advanced stages of this condition (≥1 druse of ≥125 µm). Nobenefit was observed among patients with earlier stages of age-related maculopathy(drusen <125 µm or pigmentary abnormalities without large drusen).However, the power to detect a long-term benefit in patients with early-stageAMD, considering the low rate of progression to end-stage disease in thisgroup (1.3% during 6 years) was low. The AREDS was not long enough or largeenough to rule out potential benefit to people with early age-related maculopathy.However, in recent clinical trials of high-dose vitamin E (500 IU)³³ or vitamin E (50 IU) and/or beta-carotene (20 mg),³⁴ no benefit was observed among the study subjects,most of whom had no or mild signs of AMD.²⁷

How can we generalize the results of AREDS to patients in individualpractices? People who participated in AREDS were healthier than the generalpopulation, having half the rate of mortality of a comparable general population.² The benefit could be greater or less in a populationthat is less healthy. We will be more certain about the magnitude of impactand ability to project this potential benefit across the broader populationof people with AMD when there are results from long-term prospective cohortstudies, which generally represent a wider cross section of the population.Physicians sometimes are asked whether people with a family history of AMDmight benefit from the high-dose supplements tested in AREDS. To date, thereis no evidence to support a benefit of these supplements and some evidenceof risk (discussed in the next section).

Antioxidants have been hypothesized to improve insulin sensitivity orto reduce the complications of diabetes in the eye and elsewhere. Despitesubstantial experimental and clinical evidence of high levels of oxidativestress in people with diabetes and of the involvement of oxidative stressin the development of complications, evidence of a protective effect of antioxidantshas not been demonstrated in either large-scale prospective cohort studiesor in clinical trials (reviewed by Rosen et al⁴).

Moreover, results from 3 observational studies do not suggest lowerrisk of prevalent retinopathy among people with diabetes who have higher intakesof vitamins C or E from foods and supplements^35-37 However,the changing patterns of eating and supplementation in people with diabetesmake potential protective effect hard to observe in studies, such as these,that use a cross-sectional design. In one of these studies, in which dietand supplement information was collected 6 years before retinal photographswere obtained to assess retinopathy in the Atherosclerosis Risk in CommunitiesStudy, the use of supplements containing ascorbic acid or vitamin E for 3or more years was associated with lower risk of retinopathy.³⁷ However,because of the overall unsupportive epidemiologic data at this time, the valueof antioxidant supplements in slowing or preventing diabetic retinopathy isunknown and prospective studies and clinical trials are needed to better understandpotential benefits.

The larger body of current data does not support other health benefitsof high-dose antioxidants taken individually or in combination on overallhealth.⁹ While some early studies suggestedlower rates of heart attacks among people who take vitamin E at levels higherthan 100 IU, a level only attainable from supplements,^38,39 numerousother clinical trials have shown no or limited benefit of vitamin E in supplements(reviewed by Willett and Stampfer⁹), especiallythose with high risk of heart disease. Recently, high-dose antioxidants inpeople who were also taking high doses of niacin and statin drugs resultedin an unexpected reduction in high-density lipoprotein levels, and the riskof heart attacks and strokes was increased.⁴⁰ Thepossibility that vitamin E will help, rather than hurt, people at earlierstages of atherosclerosis has not been demonstrated but is still being studied.

Oxidative stress has been hypothesized to contribute to the developmentof Alzheimer disease, as well as cancer. Studies of the benefits of antioxidantsupplement use in lowering the risk of Alzheimer disease have conflictingresults.^41-43 Similarly,clinical trials of specific antioxidants (ascorbic acid, vitamin E, and betacarotene) usually have not supported the benefit in preventing cancer, withthe exception of vitamin E in prostate cancer in male smokers and seleniumin prostate, lung, and colon cancer (reviewed by Willett and Stampfer⁹).

There are other added health risks associated with the consumption ofnutrients at these high levels. After consideration that vitamin E consumptionresults in slightly higher risk of hemorrhagic stroke, the Food and NutritionBoard set an intake of up to 1000 IU of vitamin E to be safe.²⁵ Short-term(4 months) vitamin E supplementation of less than 800 IU showed no adverseeffects on plasma lipid or lipoprotein profile, white blood cells, plateletnumbers, or bleeding time.⁴⁴ However, high-dosesupplementation may warrant caution when antithrombotic compounds such asaspirin are taken, because of a decrease in platelet adhesion and possiblebleeding (previously reviewed by Horwitt⁴⁵).

Previous studies have shown that high-dose supplemental beta carotene(≥20 mg) increased the risk of lung cancer.^46,47 Thereis currently no upper tolerable intake for beta carotene, but the DietaryReference Intake Committee of the Food and Nutrition Board does not recommendconsumption of beta carotene supplements for the general public.¹²

Clear benefit of high-dose antioxidant supplements in slowing progressionof AMD is strongly suggested by the results of the AREDS. Because of potentialrisks for other health conditions, we should be cautious in recommending supplementationat high levels. Nutrients taken in the high doses that are in such supplementshave pharmacologic benefits and risks that may not be present when consumedat the physiologic levels provided in foods. With time, it is likely thatwe will have a better understanding of which diseases high-dose antioxidantsmay most likely benefit and which may be exacerbated. An understanding ofa patient's clinical and family history may guide a physician in making recommendationsthat are tailored to the diseases a person is likely to be at risk for. Atthis time, such knowledge is limited.

However, some general guidelines for maximizing benefit and minimizingharm can be considered. If high-dose antioxidant supplements are taken, therisk may be less if several are taken concurrently. Antioxidant enzymes andmicronutrients interact with each other in a chain of events to lower oxidativestress within cells. If one element of this complex set is taken in high levels,there is potential for a perturbation of this balance and greater harm.⁴⁸ Combination supplements and concurrent intake ofa balanced diet may lower the chance for harm. Some antioxidants, like vitaminC, have been shown to be pro-oxidants in high levels or under certain conditions.⁴⁹ For these reasons, patients who are advised to takeantioxidants can be given guidance about prudent levels to consume (Table 1), the use combination preparations,rather than single nutrients, and the importance of eating a well-balanceddiet.

Zinc concentrations are high in the retinal pigment epithelium, relativeto other tissues in the body, where it may protect the eye from age-relateddamage as a cofactor for antioxidant enzymes or other enzymes that are importantto normal functioning of the retinal pigment epithelium cells (reviewed byMares-Perlman and Klein¹¹). Alternatively,zinc may protect against the effects of inflammation. A possible role of inflammationin the pathogenesis of AMD is suspected.⁵⁰

High-dose zinc supplements (80 mg/d, which is 5 to 10 times above theRDA in men and women²⁵), together with antioxidants,were shown in the AREDS to reduce the risk of progression to advanced AMDamong patients who already had extensive drusen.² Thislevel of zinc is above the recommended upper tolerable limit for zinc intake(Table 1). The zinc in this studywas provided as the zinc oxide salt. Because of the different bioavailabilityof different zinc salts that are available in supplements,⁵¹ theeffective and safe level provided by other zinc supplements (such as zincgluconate or zinc sulfate) may differ. This benefit of zinc was similar tothat provided by antioxidants alone or in combination with zinc. This resultis in agreement with one⁵² but not another⁵³ smaller and shorter-term clinical trial of zinc supplementation.Observational studies in 2,^54,55 butnot 2 other,^18,56 populationsalso support the possibility of lower risk of certain types of early age-relatedmaculopathy (ARM) among people with higher intakes of zinc. Thus, the evidencegenerally suggests that zinc supplementation may reduce the progression ofAMD in people who already have intermediate stages, but the influence on earlystages is unclear.

The influence of zinc on other aspects of health has been controversial.Studies of benefit of zinc on protection against the common cold are conflicting.⁵⁷ There are no other reports, to our knowledge, thatzinc supplements benefit overall health.

There are, however, several cautionary reports on the possible adversehealth effects of zinc supplementation. Acute zinc toxicity is fairly rare.However, excess zinc ingestion (100-300 mg/d) has been observed to resultin gastric pain, nausea, and dizziness.⁵⁸ Long-termexcess zinc supplementation may reduce copper or iron absorption.⁵⁸ Zinc supplementation of only 50 mg/d has been shownto decrease high-density lipoprotein cholesterol level,⁵⁹ butthis may be secondary to interference with copper absorption. In the AREDS,² zinc supplementation (together with copper) did notsignificantly influence blood lipid levels, but longer-term effects have notbeen studied.

Zinc supplementation resulted in more frequent hospitalizations dueto genitourinary conditions and self-reported anemia (but not hematocrit)among AREDS participants.² However, overallmortality was not higher during the 7-year period. Negative disturbances inglucose homeostasis have been documented with large-dose zinc supplementationin mice⁶⁰ and in some patients with type 1and type 2 diabetes mellitus.^61,62 Thereis also evidence to suggest that zinc might play a role in the initiationand promotion of a variety of neoplasias,⁶³ anda recent report of higher rates of advanced prostate cancer among men whotook either high doses of zinc (>100 mg) or zinc supplements for longer than10 years.⁶⁴ The relationship between zinc supplementationand diabetes and cancer has not been well studied.

Until there is more information about the safety of zinc supplements,caution is warranted. It is easy to overdose on zinc, as it is for other minerals,because the body does not have efficient ways to eliminate excesses as itdoes for water-soluble vitamins. The National Academy of Sciences²⁵ has currently set an upper limit of safety at 40mg/d for adults. This is based on the interference with copper status. Somesupplements (such as the one used in AREDS) also provide copper, which wouldreduce concern regarding the influence on copper status. However, the longer-termimpact of taking high levels of zinc in supplements and the impact in broaderpopulations of people are currently unknown. Because zinc was not a frequentcomponent of vitamin supplements before 1980, the long-term risks of supplementationwith this nutrient, particularly at high levels, have not been well studied.

The current recommended dietary intake is 8 mg/d in women and 11 mg/din men. Factors that might increase the level of zinc needed in a supplementare vegetarian diets, intake of iron supplements, and heavy alcohol intake(previously reviewed²⁵). However, for mostpeople, it is currently prudent to get adequate zinc by eating meats and drinkingmilk. In general, zinc bioavailability from a diet that is high in animalprotein, as compared with vegetable protein (such as soy), is high.²⁴ Standard multivitamins with minerals provide zinc(15 mg/d) at levels that are commonly provided in foods.

Recent attention has focused on dietary components that are not essentialnutrients, but may otherwise have beneficial health effects. The potentialfor the plant pigments lutein and its isomer zeaxanthin to slow the developmentof age-related eye disease is under investigation. These are plant pigmentsfrom the carotenoid family, like beta carotene. Unlike beta carotene, theycannot form vitamin A. Their importance to eye tissues is suggested by thefact that they are the only carotenoids in the eye, unlike all other bodytissues, which contain a wide variety of carotenoids such as beta carotene.This may be due to their abilities to absorb blue light, to function as antioxidants,or both. Evidence (reviewed by Mares-Perlman et al⁶⁵)suggests that lutein and zeaxanthin may reduce the risk of developing the2 most common eye diseases in older people, cataract and macular degeneration.Currently, however, the evidence is far from conclusive. There are many piecesof evidence still to be evaluated and inconsistencies across studies. Thereis also the untested possibility that lutein and/or zeaxanthin may improvevision or slow progression once these conditions are present.

The intake of lutein and zeaxanthin has been associated with lower riskof cardiovascular disease or cancers in some previous studies (reviewed byMares-Perlman et al⁶⁵). However, there arefew studies that evaluate the association with overall health. Also, as isthe case with eye diseases, there is no experimental proof that lutein andzeaxanthin benefit health.

Long-term risk of taking lutein supplements is unstudied. For this reason,it is prudent to rely on foods to get these pigments or consider taking levelsin supplements that are commonly consumed from foods. The average intake inAmericans is 1.3 mg/d; 10% of Americans consumed more than 6 mg/d.⁶⁶ Because lutein can compete with other carotenoidsfor absorption⁶⁷ and because harmful effectsof other similar carotenoids (beta carotene, as discussed earlier) have beenobserved, taking higher levels is not recommended before more is known.

Herbal medicines have long been used in some cultures to improve eyesight.Today, a wide variety of herbal supplements sit aside more traditional nutrientmixtures in the "eye supplement" section of drug stores. Some preparationson the market combine high-dose antioxidant mixtures with herbal ingredients.

One such herbal supplement is bilberry, a species of berry native tonorthern Europe, Asia, and North America that is closely related to blueberries.These contain anthocyanin pigments that give the characteristic color of blueberries,flavonoids, and a variety of other phenolic compounds. The popularity of bilberryin Japan and Korea relates to the belief that it improves night vision. Thismay be due to the carotenoid content of bilberries, which contain vitaminA activity, or due to independent influences of anthocyanin, for which thereis some evidence of effectiveness but limited to small studies (reviewed byCamire⁶⁸). The effectiveness and safety ofsuch herbal supplements are highly discussed in marketing literature but remainunsubstantiated in the scientific literature.

Considering the evidence discussed in this article, we offer the followingrecommendations for clinicians.

Current evidence supports the benefit of specific supplements (thatcontain 500 mg of ascorbic acid, 400 IU of vitamin E, and 80 mg of zinc with2 mg of copper) and beta carotene (25 000 IU) in patients who alreadyhave intermediate stages of AMD in slowing progression to end-stage disease.The short-term safety of this supplement is suggested in healthier-than-normalpeople who participated in this trial. However, longer-term safety is unknown.For these reasons, the following cautionary notes are made. Patients who maybe at high risk for adverse effects should be monitored, including (1) patientswith diabetes (for possible glucose intolerance with zinc), (2) patients whotake aspirin or have a history of hemorrhage or easy bruising (there is evidencethat the high dose of vitamin E may contribute to a tendency to bleed), and(3) patients who are currently using statins to lower blood lipid levels;this formulation might, because of the high vitamin E levels, lower high-densitylipoprotein cholesterol levels and the cholesterol-lowering response to statins.Smokers should be advised to take formulations without beta carotene. Patientsshould be advised to discontinue any other supplements that contain thesenutrients except a multivitamin that contains levels that are needed dailyfrom foods (the RDAs).

Because there are likely to be other benefits from food components thatare not contained in supplements that we do not yet understand, such as thebenefit of the plant pigments lutein and zeaxanthin and possibly vitamin Ecompounds that are not contained in supplements, patients should also be advisedto follow the US dietary guidelines for healthy eating.⁶⁹ Theseguidelines include, among other things, getting 3 to 5 servings of vegetablesand 2 to 4 servings of fruits, 2 or 3 servings of dairy products, and 2 or3 servings of meats, fish, eggs, or beans daily. If getting the adequate numberof servings of fruits and vegetables and including among them good sourcesof lutein (dark-green vegetables) is not likely, a lutein supplement thatcontains levels that are likely to come from foods (1-6 mg) is advisable whileresearch on the specific benefits and risks of consuming these carotenoidscontinues.

Currently, there is no evidence that taking supplements prevents theearly stages of ARM or the development of diabetic retinopathy. Patients whohave a family history of these conditions should weigh the many potentialadverse effects of supplement use against the unproven benefit. Getting adequatenutrients and other food components in a healthy diet (as discussed earlier)may be the best and safest insurance against nutrient inadequacy contributingto risk of ARM in the future. If patients decide to take supplements, riskis minimized by (1) taking multivitamin or combination supplements ratherthan single supplements and at levels that are close to the RDA and do notexceed the upper tolerable limit set by the Food and Nutrition Board (Table 1); and (2) avoiding excessive vitaminA intake in supplements by taking supplements that provide vitamin A as betacarotene, rather than preformed retinol, or avoiding concurrent use of vitamin-fortifiedcereals, high dietary intake of vitamin A (meats and milk), and liquid meal-replacementsupplements (Table 2).

Taking multivitamins may slow the development of age-related cataracts,but because this is unproven, patients should consider taking multivitaminson the basis of overall health benefits (providing nutrients to meet dailyneeds because of inability to eat a healthful diet) or risks (the possibilityof hip fracture risk being increased if vitamin A intake from all sourcescombined is high, as discussed previously). Diets that meet the dietary guidelinesmay be the safest way to obtain adequate nutrients and have the added benefitof providing nonnutrient diet components such as lutein and zeaxanthin thatmay also lower risk of cataract.

Corresponding author and reprints: Julie A. Mares, PhD, Departmentof Ophthalmology and Visual Sciences, University of Wisconsin–Madison,610 N Walnut St, 1062 WARF, Madison, WI 53705-9279 (e-mail: jmarespe@facstaff.wisc.edu).

Submitted for publication November 13, 2002; final revision receivedJuly 25, 2003; accepted August 21, 2003.

This study was supported by grant EY 13018 from the National Eye Institute,National Institutes of Health, Bethesda, Md, and by Research to Prevent BlindnessInc, New York, NY.