HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Carr, A. C Articles by Frei, B. Search for Related Content PubMed PubMed Citation Articles by Carr, A. C Articles by Frei, B. Agricola Articles by Carr, A. C Articles by Frei, B.

Toward a new recommended dietary allowance for vitamin C based on antioxidant and health effects in humans^1,2,3

	ABSTRACT

TOP ABSTRACT INTRODUCTION VITAMIN C AS AN... SPECIAL POPULATIONS SUMMARY AND CONCLUSIONS REFERENCES

 
The current recommended dietary allowance (RDA) for vitamin C for adult nonsmoking men and women is 60 mg/d, which is based on a mean requirement of 46 mg/d to prevent the deficiency disease scurvy. However, recent scientific evidence indicates that an increased intake of vitamin C is associated with a reduced risk of chronic diseases such as cancer, cardiovascular disease, and cataract, probably through antioxidant mechanisms. It is likely that the amount of vitamin C required to prevent scurvy is not sufficient to optimally protect against these diseases. Because the RDA is defined as "the average daily dietary intake level that is sufficient to meet the nutrient requirement of nearly all healthy individuals in a group," it is appropriate to reevaluate the RDA for vitamin C. Therefore, we reviewed the biochemical, clinical, and epidemiologic evidence to date for a role of vitamin C in chronic disease prevention. The totality of the reviewed data suggests that an intake of 90–100 mg vitamin C/d is required for optimum reduction of chronic disease risk in nonsmoking men and women. This amount is about twice the amount on which the current RDA for vitamin C is based, suggesting a new RDA of 120 mg vitamin C/d.

Key Words: Antioxidant • cancer • cardiovascular disease • cataract • DNA • lipid • protein • recommended dietary allowance • vitamin C • adults

	INTRODUCTION

TOP ABSTRACT INTRODUCTION VITAMIN C AS AN... SPECIAL POPULATIONS SUMMARY AND CONCLUSIONS REFERENCES

 
Vitamin C (ascorbic acid) is an essential micronutrient required for normal metabolic functioning of the body (1). Humans and other primates have lost the ability to synthesize vitamin C as a result of a mutation in the gene coding for L-gulonolactone oxidase, an enzyme required for the biosynthesis of vitamin C via the glucuronic acid pathway (2). Thus, vitamin C must be obtained through the diet. The vitamin is especially plentiful in fresh fruit, in particular citrus fruit, and vegetables (3). A lack of vitamin C in the diet causes the deficiency disease scurvy (4). This potentially fatal disease can be prevented with as little as 10 mg vitamin C/d (5), an amount easily obtained through consumption of fresh fruit and vegetables.

The current recommended dietary allowance (RDA) for vitamin C is 60 mg/d for healthy, nonsmoking adults (6). The RDA is determined by the rate of turnover and rate of depletion of an initial body pool of 1500 mg vitamin C and an assumed absorption of 85% of the vitamin at usual intakes (7). This amount provides an adequate margin of safety: 60 mg/d would prevent the development of scurvy for 1 mo with a diet lacking vitamin C (7). The RDAs are determined primarily on the basis of prevention of deficiency; because scurvy is not a major health problem in the United States, this goal is clearly accomplished by the current RDA for vitamin C. Nevertheless, 25% of men and women in the United States consume <60 mg vitamin C/d and 10% of adults consume <30 mg/d (3).

The molecular mechanisms of the antiscorbutic effect of vitamin C are largely, although not completely, understood. Vitamin C is a cofactor for several enzymes involved in the biosynthesis of collagen, carnitine, and neurotransmitters (8, 9). Procollagen-proline dioxygenase (proline hydroxylase) and procollagen-lysine 5-dioxygenase (lysine hydroxylase), 2 enzymes involved in procollagen biosynthesis, require vitamin C for maximal activity (10). Posttranslational hydroxylation of proline and lysine residues by these enzymes is essential for the formation and secretion of stable collagen helixes. A deficiency of vitamin C results in a weakening of collagenous structures, causing tooth loss, joint pains, bone and connective tissue disorders, and poor wound healing, all of which are characteristics of scurvy (8). Two dioxygenases involved in the biosynthesis of carnitine also require vitamin C as a cofactor for maximal activity (8). Carnitine is essential for the transport of activated long-chain fatty acids into the mitochondria; as a result, vitamin C deficiency results in fatigue and lethargy, early symptoms of scurvy. In addition, vitamin C is used as a cofactor for catecholamine biosynthesis, in particular the conversion of dopamine to norepinephrine catalyzed by dopamine ß-monooxygenase (8). Depression, hypochondria, and mood changes frequently occur during scurvy and could be related to deficient dopamine hydroxylation.

The activities of several other enzymes are known to be dependent on vitamin C, although their connection to scurvy has not yet been clearly established. These enzymes include the mono- and dioxygenases involved in peptide amidation and tyrosine metabolism (8, 9). Vitamin C has also been implicated in the metabolism of cholesterol to bile acids via the enzyme cholesterol 7-monooxygenase and in steroid metabolism in the adrenals (8, 9). Hydroxylation of aromatic drugs and carcinogens by hepatic cytochrome P450 is also enhanced by reducing agents such as vitamin C (9).

The role of vitamin C in the above metabolic pathways is to reduce the active center metal ion of the various mono- and dioxygenases (8, 9). Unlike other water-soluble vitamins, vitamin C acts as a cosubstrate in these reactions, not as a coenzyme. The ability to maintain metal ions in a reduced state is related to the redox potential of vitamin C (11). Reduction of iron by vitamin C has also been implicated in enhanced gastrointestinal absorption of dietary nonheme iron (8, 12). Other proposed activities of vitamin C include maintenance of enzyme thiols in a reduced state and sparing of glutathione, an important intracellular antioxidant and enzyme cofactor (13), and tetrahydrofolate, a cofactor required for catecholamine biosynthesis (9).

Many biochemical, clinical, and epidemiologic studies have indicated that vitamin C may be of benefit in chronic diseases such as cardiovascular disease, cancer, and cataract (5, 14, 15). The amount of vitamin C required to prevent scurvy may be less than the amount necessary to maintain optimal health and reduce the incidence of chronic disease morbidity and mortality. The Food and Nutrition Board of the US National Academy of Sciences has changed the criteria for establishing RDAs from prevention of deficiency disease to prevention of chronic diseases. Therefore, in this review we address whether the current RDA of 60 mg vitamin C/d is sufficient for optimally reducing the risk of chronic diseases such as cardiovascular disease and cancer, as well as cataract. We also address whether the activity of vitamin C in these conditions is due to its antioxidant properties or to other proposed mechanisms.

	VITAMIN C AS AN ANTIOXIDANT

TOP ABSTRACT INTRODUCTION VITAMIN C AS AN... SPECIAL POPULATIONS SUMMARY AND CONCLUSIONS REFERENCES

 
Vitamin C is an important water-soluble antioxidant in biological fluids (16, 17). An antioxidant has been defined as "any substance that, when present at low concentrations compared to those of an oxidizable substrate (e.g., proteins, lipids, carbohydrates and nucleic acids), significantly delays or prevents oxidation of that substrate" (11). The definition proposed by the Panel on Dietary Antioxidants and Related Compounds of the Food and Nutrition Board is that "a dietary antioxidant is a substance in foods that significantly decreases the adverse effects of reactive oxygen species, reactive nitrogen species, or both on normal physiological function in humans" (18). Vitamin C readily scavenges reactive oxygen and nitrogen species, such as superoxide and hydroperoxyl radicals, aqueous peroxyl radicals, singlet oxygen, ozone, peroxynitrite, nitrogen dioxide, nitroxide radicals, and hypochlorous acid (11), thereby effectively protecting other substrates from oxidative damage. Although vitamin C also reacts rapidly with hydroxyl radicals (rate constant >10⁹ L • mol^-1•s^-1), it is nevertheless unable to preferentially scavenge this radical over other substrates (19). The reason for this is that hydroxyl radicals are extremely reactive and will combine indiscriminately with any substrate in their immediate environment at a diffusion-limited rate. Vitamin C can also act as a coantioxidant by regenerating -tocopherol (vitamin E) from the -tocopheroxyl radical, produced via scavenging of lipid-soluble radicals (20, 21). This is a potentially important function because in vitro experiments have shown that -tocopherol can act as a prooxidant in the absence of coantioxidants such as vitamin C (21, 22). However, the in vivo relevance of the interaction between vita- min C and vitamin E is unclear. Vitamin C has also been shown to regenerate urate, glutathione, and ß-carotene in vitro from their respective one-electron oxidation products, ie, urate radicals, glutathiyl radicals, and ß-carotene radical cations (11, 23).

Two major properties of vitamin C make it an ideal antioxidant. First is the low one-electron reduction potentials of both ascorbate (282 mV) and its one-electron oxidation product, the ascorbyl radical (-174 mV), which is derived from the ene-diol functional group in the molecule (11). These low reduction potentials enable ascorbate and the ascorbyl radical to react with and reduce basically all physiologically relevant radicals and oxidants. For this reason, vitamin C has been said to be "at the bottom of the pecking order" and "to act as the terminal water-soluble small molecule antioxidant" in biological systems (24). The second major property that makes vitamin C such an effective antioxidant is the stability and low reactivity of the ascorbyl radical formed when ascorbate scavenges a reactive oxygen or nitrogen species (Eq 1). The ascorbyl radical readily dismutates to form ascorbate and dehydroascorbic acid (Eqs 1 and 2), or is reduced back to ascorbate by an NADH-dependent semidehydroascorbate reductase (9, 20, 25). The 2-electron oxidation product of ascorbate, dehydroascorbic acid, can itself be reduced back to ascorbate by glutathione, the glutathione-dependent enzyme glutathione:dehydroascorbate oxidoreductase [glutathione dehydrogenase (ascorbate), or glutaredoxin], or the NADPH-dependent selenoenzyme thioredoxin reductase (9, 20, 25). Alternatively, dehydroascorbic acid is rapidly and irreversibly hydrolyzed to 2,3-diketogulonic acid (DKG) (Eq 3) (11).

(1)

(2)

(3)

where equation 1 shows the reversible 1- and 2-electron oxidation of ascorbate (AH^-) to the ascorbyl radical (A^•-) and dehydroascorbic acid (A), respectively; equation 2 shows the dismutation of the ascorbyl radical to form ascorbate and dehydroascorbic acid; and equation 3 shows the hydrolysis of dehydroascorbic acid to DKG, which then decomposes to oxalate, threonate, and many other products.

Vitamin C has been recognized and accepted by the US Food and Drug Administration (FDA) as one of 4 dietary antioxidants, the other 3 being vitamin E, the vitamin A precursor ß-carotene, and selenium, an essential component of the antioxidant enzymes glutathione peroxidase and thioredoxin reductase. The Panel on Dietary Antioxidants and Related Compounds of the Food and Nutrition Board (26) has, in principle, concurred with this definition, and in addition will consider other carotenoids. New regulations were recently published in which the FDA stated that vitamin C serves as an effective free radical scavenger to protect cells from damage by reactive oxygen molecules (27). Statements of the antioxidant properties of vitamin C are also appearing on food labels and in nutrient content claims throughout the United States.

Although substantial scientific evidence exists regarding the antioxidant and health effects of vitamin C in humans, further investigations of the role of vitamin C both in vitro and in vivo are warranted, particularly because vitamin C, being a redox-active compound, can act not only as an antioxidant, but also as a prooxidant in the presence of redox-active transition metal ions (11). Reduction of metal ions such as iron or copper by vitamin C in vitro (Eq 4) can result in the formation of highly reactive hydroxyl radicals via reaction of the reduced metal ions with hydrogen peroxide, a process known as Fenton chemistry (Eq 5). Lipid hydroperoxides may also be broken down by the reduced metal ions, resulting in the formation of lipid alkoxyl radicals (Eq 6) that can initiate and propagate the chain reactions of lipid peroxidation (28). The mechanism shown in equation 5, however, requires the availability of free, redox-active metal ions and a low ratio of vitamin C to metal ion, conditions unlikely to occur in vivo under normal circumstances (11, 15, 28). Furthermore, it was shown recently that in biological fluids such as plasma, vita-min C acts as an antioxidant toward lipids even in the presence of free, redox-active iron (29).

(4)

(5)

(6)

where equation 4 shows the reduction of redox-active metal ions [M⁽ⁿ⁺¹⁾] by ascorbate to form the ascorbyl radical and the reduced metal (Mⁿ), equation 5 shows the production of highly reactive hydroxyl radicals (^•OH) from the reaction of hydrogen peroxide (H₂O₂) with the reduced metal ions, and equation 6 shows the reaction of lipid hydroperoxides (LOOH) with reduced metal ions to form alkoxyl radicals (LO^•).

Although there is no convincing evidence for a prooxidant effect of vitamin C in humans, there is substantial evidence for vitamin C's antioxidant activity. Interestingly, the antioxidant activity of vitamin C is not directly related to its antiscorbutic effect. Therefore, if the antioxidant activity of vitamin C is accepted as occurring in vivo and considered to be relevant to human health by the Panel on Dietary Antioxidants and Related Compounds of the Food and Nutrition Board, then scurvy should not be used as the sole criterion for nutritional adequacy or to determine the required or optimal amount of vitamin C. In this section of the review, therefore, we address the following questions: 1) Is oxidative damage to biological macromolecules relevant to human chronic diseases? 2) What is the evidence that vita-min C acts as an antioxidant in humans? 3) Are there other mechanisms, besides those directly related to the antiscorbutic and antioxidant activities of vitamin C, by which vitamin C could affect chronic disease incidence, mortality, or both? and 4) Does vitamin C lower chronic disease incidence, mortality, or both?

Is oxidative damage to biological macromolecules relevant to human chronic diseases?
Oxidative damage to biomolecules, such as lipids, DNA, and proteins, has been implicated in many chronic diseases, in particular, cardiovascular disease, cancer, and cataract, respectively (30).

LDL oxidation and cardiovascular disease
Oxidative processes have been strongly implicated in atherosclerosis, myocardial infarction, and stroke (31). The oxidative modification hypothesis of atherosclerosis is currently the most widely accepted model of atherogenesis. LDL, the major carrier of cholesterol and lipids in the blood (32), infiltrates the intima of lesion-prone arterial sites, where it is oxidized over time by oxidants generated by local vascular cells or enzymes (33) to a form that exhibits atherogenic properties. Minimally oxidized LDL and cytokines can activate endothelial cells to express surface adhesion molecules, primarily vascular cell adhesion molecule 1 and intercellular adhesion molecule 1 as well as monocyte chemotactic protein 1, which cause circulating monocytes to adhere to the endothelium and migrate into the artery wall (31). The monocytes subsequently differentiate into macrophages in response to macrophage colony stimulating factor, the expression of which by vascular cells also is enhanced by modified LDL. The oxidized LDL further inhibits the egress of macrophages from the artery wall, where the cells recognize and readily take up the oxidized LDL through a process mediated by scavenger receptors (31). Unlike the normal apolipoprotein B/E LDL receptor recognizing native LDL, the scavenger receptors on macrophages that recognize modified LDL are not tightly regulated; as a result, the macrophages are converted into foam cells, a component of fatty streaks and the hallmark of atherosclerosis.

It is still uncertain which factors are responsible for the oxidation of LDL in vivo. LDL can be oxidized into a potentially atherogenic form in vitro through metal-ion-dependent oxidation of its lipid component with subsequent modification of apolipoprotein B-100 by reactive aldehyde products of lipid peroxidation, particularly malondialdehyde and 4-hydroxynonenal (34). Whether catalytic metal ions are available in the early lesion in vivo remains a matter of debate (28). Several metal-ion-independent mechanisms that are primarily enzymatic in nature have been proposed; these include mechanisms involving 15-lipoxygenase and myeloperoxidase (33, 35). The problems of comparing LDL oxidation in vitro with LDL oxidation in vivo were discussed in 2 recent reviews (36, 37).

Several lines of direct evidence point to the formation and existence of oxidized LDL in vivo. Antibodies to aldehyde-modified LDL recognize epitopes in human plaques (38) and LDL extracted from these lesions reacts with antibodies to oxidized LDL and has characteristics identical to those of LDL oxidized in vitro. Aldehyde-modified LDL has also been detected in plasma, as have autoantibodies to oxidized LDL (31). Antibodies to hypochlorous acid–modified protein have detected epitopes in lesions, suggesting an alternative or additional mechanism of LDL oxidation (39). F₂-isoprostanes, stable markers of lipid oxidation, have been detected in lesions (40, 41). Other oxidized lipids also increase with age and severity of atherosclerosis (42). Indirect evidence for in vivo oxidation has come from antioxidant supplementation studies in animals that show reduced lesion formation and reduced LDL oxidation (43, 44). In addition, numerous epidemiologic studies have indicated that dietary antioxidants reduce the incidence of cardiovascular disease in humans, as discussed below.

DNA oxidation and mutagenesis and carcinogenesis
Carcinogenesis is a multistage process. Free radicals and oxidative processes have been implicated in both the initiation and the promotion of carcinogenesis (5, 45). The oxidative hypothesis of carcinogenesis asserts that many carcinogens can generate free radicals that damage cells, predisposing these cells to malignant changes (46). Antioxidants, by neutralizing free radicals and oxidants, can prevent cell damage and subsequent development of cancer. DNA contains reactive groups in its bases that are highly susceptible to free radical attack (47) and oxidative DNA damage can lead to deleterious mutations (2). It has been proposed that oxidative damage to DNA occurs continuously in vivo at a rate of 10⁴ oxidative hits per cell per day (2). Most oxidative lesions are efficiently repaired by specific DNA glycosylases, but repair is not 100% efficient and the number of lesions accumulates with age. When the cells divide, the lesions become fixated and mutations and cancer may result (2). The existence of numerous DNA glycosylases specific for oxidized bases (47) indicates that the latter are deleterious to human survival and therefore must be repaired.

There are >20 different oxidative DNA lesions, of which 8-oxoguanine is one of the most abundant and best studied (2). 8-Oxoguanine is formed by hydroxyl radical, peroxynitrite, or singlet oxygen attack of guanine in the C-8 position and is highly mutagenic (2). Normally guanine forms a base pair with cytosine, but 8-oxoguanine tends to pair with adenine. After replication, a transverse mutation occurs whereby G-C base pairs are replaced by A-T base pairs. This may contribute to the formation of cancer if the lesion occurs at a critical position in an important gene, such as a tumor suppressor or growth factor gene (2). Smokers, who have a higher risk of lung cancer than nonsmokers, have elevated concentrations of 8-oxo-2'-deoxyguanosine (8-oxodG) (48–50). 8-Oxoguanine has also been detected in women with breast cancer (51, 52). Although direct evidence for the link between DNA oxidation and cancer is still lacking (53), many epidemiologic studies have suggested that dietary intake of antioxidant vitamins, mainly from fruit and vegetables, protects against different types of cancer.

Protein oxidation and cataract
Cataract is a dysfunction of the lens resulting from opacification, which impedes the transmission of light (54). About 98% of the solid mass of the lens is protein, predominantly crystallins. These proteins are long lived and undergo minimal turnover; as a result, cataract formation is primarily age related. Oxidation of the lens proteins as a result of chronic exposure to ultraviolet light and oxygen has been implicated in this process (54, 55). Smoking, which is known to produce oxidative stress, is also associated with enhanced cataract risk (56, 57). Evidence of lens protein oxidation includes loss of sulfhydryl and tryptophan residues with age as well as formation of disulfides and other covalent cross-linkages. Deamination and acidification also occur, as well as formation of advanced glycation end products, particularly in persons with diabetes (54, 58). The oxidized proteins accumulate, aggregate, and eventually precipitate, producing the sequelae of cataract.

The lens contains multiple antioxidant defenses, such as high concentrations of vitamin C and glutathione, and antioxidant enzymes such as superoxide dismutase, catalase, and the glutathione peroxidase-reductase system (54, 55). Secondary defenses include proteolytic enzymes that selectively degrade damaged proteins. With aging, however, antioxidant concentrations in the lens, including concentrations of vitamin C, may be reduced (54) and the antioxidant enzymes are prone to inactivation, resulting in increased protein oxidation in older lenses. Proteolytic activity is also reduced in older lenses, resulting in accumulation of damaged proteins (54). Therefore, supplementation with antioxidants such as vitamin C may reduce the risk of cataract.

What is the evidence that vitamin C acts as an antioxidant in humans?
The most conclusive evidence that vitamin C acts as an antioxidant in humans has come from supplementation studies using specific biomarkers of oxidative damage to lipids, DNA, and proteins. Because these specific oxidative biomarkers have only recently been developed and continue to be evaluated, only relatively few studies have investigated the effects on these biomarkers of supplementation with antioxidant micronutrients, including vitamin C.

Lipid oxidation
The most commonly used assay for lipid oxidation, although perhaps the least specific, measures the aldehyde peroxidation product malondialdehyde and other aldehydes by reaction with thiobarbituric acid (the so-called thiobarbituric acid–reactive substances, or TBARS, assay) (59, 60). Other products of lipid peroxidation, such as conjugated dienes and lipid hydroperoxides, are often measured (61, 62). TBARS and conjugated dienes are also commonly used to measure the oxidizability, or susceptibility to oxidation, of LDL (63). The oxidizability of LDL is determined by measuring the lag time and propagation rate of lipid peroxidation in LDL exposed in vitro to copper ions or other oxidants and is dependent on the antioxidant content and lipid composition of LDL (32, 63). Specific biomarkers of lipid peroxidation are the F₂-isoprostanes [in particular 8-epi-prostaglandin F₂ (8-epi-PGF₂)], which are formed from nonenzymatic, radical-mediated oxidation of arachidonyl-containing lipids (59). Increased concentrations of F₂-isoprostanes have been detected in persons with diabetes, in smokers, in persons with hypercholesterolemia (64, 65), and in LDL exposed in vitro to various types of oxidative stress (66).

Numerous studies in humans have investigated the effects on the oxidizability of LDL of vitamin C supplementation in combination with vitamin E or ß-carotene or both (67). Studies have been carried out in smokers (68–70), nonsmokers (71, 72), and persons with hypercholesterolemia or cardiovascular disease (73, 74). In all cases, a significant reduction in LDL oxidizability was observed. It is, however, difficult to determine the relative contribution of vitamin C in these studies because of the presence of the cosupplements, of which vitamin E appears to be the major contributor to protection of LDL. This is because vita-min E is the most abundant lipid-soluble antioxidant associated with LDL (32).

Several studies of LDL oxidizability have also been carried out with vitamin C as the only supplement (Table 1). Although 2 studies found no effects (80, 81), 2 other studies (76, 84) found a significant reduction in the oxidizability of LDL obtained from persons supplemented with vitamin C. It is difficult to rationalize these findings, however, because vitamin C, being water soluble, is removed from the LDL during isolation from plasma. One possible explanation is the postulated role of vitamin C in either sparing or recycling vitamin E (15, 84). As mentioned above, this activity is readily observed in vitro, but evidence for sparing or recycling of vitamin E by vitamin C in vivo is inconclusive (15, 83, 87).

An equal number of studies with vitamin C supplementation have been carried out in healthy individuals or nonsmokers. A reduction in lipid oxidation was observed with some markers (83–85), whereas a nonsignificant decrease (82) or no change was observed with others (79, 83). Finally, a recent study in coronary artery disease patients supplemented with 500 mg vita-min C/d for 1 mo found no change in plasma 8-epi-PGF₂ concentrations (86). With one exception (80), none of the studies listed in Table 1 found an increase in lipid oxidation markers with vitamin C supplementation.

An antioxidant role of vitamin C in vivo is also suggested by animal studies in guinea pigs or genetically scorbutic (Osteogenic Disorder Shionogi, or ODS) rats, which, like humans, are unable to synthesize vitamin C (90–93). Guinea pigs fed a diet marginally deficient in vitamin C have significantly higher concentrations of endogenous malondialdehyde than do animals fed diets supplying 20–40 times the amount of vitamin C required to avoid scurvy (90). Vitamin C–deficient guinea pigs exhale significantly higher amounts of breath ethane when exposed to an oxidative stress than do vitamin C–sufficient animals (92). Note that human smokers supplemented with 600–1000 mg vitamin C/d in combination with vitamin E and ß-carotene for 3–8 wk exhaled lower amounts of breath pentane and ethane (68, 94). Genetically scorbutic ODS rats fed a vitamin C–free diet have higher concentrations of plasma and liver TBARS and lipid peroxides than do rats fed vitamin C–supplemented diets (91). In a more recent study (93), however, supplementation of ODS rats with vitamin C, in contrast with vitamin E, did not protect against the endogenous formation of TBARS.

Finally, in vitro experiments with human plasma have shown that the formation of F₂-isoprostanes and lipid hydroperoxides by aqueous peroxyl radicals, stimulated neutrophils, gas-phase cigarette smoke, or redox-active iron does not occur until most or all of the endogenous vitamin C has been depleted (29, 66, 95). Vitamin C has also been shown to protect isolated LDL in vitro from oxidation by various radicals and oxidants (31). It is not surprising that vitamin C effectively prevents oxidation of isolated LDL and lipoproteins in plasma because, as explained above, vitamin C effectively scavenges most aqueous reactive oxygen and nitrogen species before they can interact with and oxidize other substrates, including lipids.

DNA oxidation
Of the >20 known oxidative DNA lesions, the most commonly measured markers of in vivo DNA oxidation are 8-oxoguanine and its respective nucleoside 8-oxodG (96). These modified bases have been detected in cells such as lymphocytes and are also excreted in urine (96). The tissue pool represents a steady state between oxidation and cellular repair mechanisms, whereas the excreted products represent the total net damage and repair. The 2 methods most commonly used to detect 8-oxoguanine and 8-oxodG are gas chromatography–mass spectroscopy (GC-MS) and HPLC with electrochemical detection (HPLC-EC) (97). These methods, however, can generate artificially high amounts of oxidation products because of lengthy extraction, hydrolysis, and derivatization procedures, especially with GC-MS (97). Indirect methods for measuring DNA damage by using specific repair endonucleases in combination with single-cell gel electrophoresis (the comet assay) have shown baseline concentrations of DNA oxidation products up to 1000-fold lower than those measured by GC-MS (96, 97). Antibodies against 8-oxopurines are another potentially useful method for detecting DNA oxidation (98).

DNA oxidation, as determined by 8-oxodG in cells, is increased in cases of oxidative stress such as smoking and is correlated with reduced plasma concentrations of the antioxidant vitamins C and E (99). It is therefore conceivable that supplementation with vita-min C may ameliorate oxidative damage to DNA. Six of 7 studies shown in Table 2 (82, 100–102, 104, 105) represent cell measurements, ie, steady state damage. Five of these studies showed a significant reduction in 1 marker of oxidative DNA damage in vitamin C–supplemented subjects (100–102, 104, 105).

Another recent study using GC-MS showed a significant reduction in both 8-oxoguanine and 8-oxoadenine in healthy persons supplemented for 12 wk with either 60 or 260 mg vitamin C/d in combination with 14 mg Fe/d, although other modified bases were increased, including 5-hydroxycytosine and thymine glycol (105). Once again, however, baseline concentrations of 8-oxoguanine and 8-oxoadenine were high and comparable with those measured by Podmore et al (104). Furthermore, there were no control groups included given iron alone or placebo, and vitamin C supplementation in the study did not always result in significant changes in plasma vitamin C concentrations (105). Therefore, the interpretation of the data from these 2 studies (104, 105) is uncertain.

The study by Anderson et al (82) indicated no significant change in DNA damage as assessed by the comet assay after 2 wk of supplementation with 60 or 6000 mg vitamin C/d, either before or after an ex vivo challenge by hydrogen peroxide (Table 2). In contrast, 2 other studies in which the comet assay was used showed reduced susceptibility of lymphocytes to ex vivo oxidation of DNA after supplementation with vitamin C (101, 102). Similar findings were reported in another study in which vitamin E and ß-carotene were given as cosupplements (108). Using HPLC-EC, Fraga et al (100) also showed a significant decrease in sperm 8-oxodG after replenishment with vitamin C. A recent animal study of vitamin C (and vita- min E) supplementation of guinea pigs, however, showed no effect on 8-oxodG concentrations in the liver as determined by HPLC-EC (109), despite a 60-fold difference in liver vitamin C concentrations.

The remaining study in Table 2 (103), which represents urinary measurements (ie, the net rate of damage and repair), showed no significant effect of vitamin C supplementation on 8-oxodG concentrations as determined by HPLC-EC. Two other studies in which vitamin C was given in the presence of cosupplements also reported similar findings (49, 110). However, several investigators have questioned the appropriateness of using urinary concentrations of 8-oxoguanosine or 8-oxodG as markers of nucleic acid damage because these products can be derived from nonspecific degradation of RNA or DNA, respectively, from dead cells (47), and the major 8-oxoguanine DNA glycosylase repair product is 8-oxoguanine, not 8-oxodG.

Protein oxidation
Protein oxidation is most commonly measured by determining carbonyl groups, oxidized amino acids, and advanced glycation end products (111, 112). Protein carbonyls can be formed by direct oxidative cleavage of the peptide chain or by oxidation of specific amino acid residues, such as lysine, arginine, proline, and threonine (111). Carbonyls can also be formed indirectly through modification of lysine, histidine, and cysteine residues by ,ß-unsaturated aldehydes such as 4-hydroxynonenal in a process called Michael addition, or via Schiff base formation between lysine residues and dialdehydes such as malondialdehyde (111, 113). A variety of oxidized amino acids have been identified in proteins exposed to oxidizing conditions, including methionine sulfoxide, o- and m-tyrosine, o,o'-dityrosine, 3-chlorotyrosine, and 3-nitrotyrosine (112). Advanced glycation end products are generated by reaction of reducing sugars with lysine residues under oxidizing conditions in the Maillard reaction to form carboxymethyl- and carboxyethyl-lysine (111). Advanced glycation end products have been implicated in diseases such as diabetes and in aging (58, 111). The oxidative products of vitamin C may undergo similar reactions, although whether these occur in vivo is uncertain (114).

Few studies have been carried out to investigate the effects of vitamin C supplementation on the in vivo products of protein oxidation. Two animal studies (90, 115) indicated reduced protein carbonyl formation in guinea pigs supplemented with vitamin C, either with (115) or without (90) an endotoxin challenge. Vitamin C supplementation (2000 mg/d for 4–12 mo) of patients with Helicobacter pylori gastritis led to a significant reduction in nitrotyrosine concentrations (116). However, another recent human study found no change in urine o,o'-dityrosine or o-tyrosine concentrations in coronary artery disease patients supplemented with 500 mg vitamin C/d for 1 mo (86). Further studies investigating the effect of vitamin C supplementation on the markers of protein oxidation are clearly required.

Taken together, the evidence reviewed above suggests that vita- min C acts as an antioxidant in humans. Biomarker studies showed reduced lipid oxidation after vitamin C supplementation in both smokers and nonsmokers (Table 1). Several studies also showed decreased steady state DNA oxidation after vitamin C supplementation, although other studies showed either no change or mixed results (Table 2). The latter appears to be primarily due to the technical difficulties of accurately measuring DNA oxidation products by GC-MS and HPLC-EC and eliminating ex vivo artifacts. In addition, the effects of vitamin C supplementation on oxidative markers also critically depend on baseline concentrations of the vitamin. In a study by Levine et al (117) of the pharmacokinetics of vitamin C, it was found that in healthy adult men tissue saturation occurred at vitamin C intakes of 100 mg/d, as assessed by vitamin C concentrations in lymphocytes, monocytes, and neutrophils. Kallner et al (118) also reported that the body pool of vitamin C was saturated by an intake of 100 mg/d in healthy, nonsmoking men. If tissues are already saturated before vitamin C supplementation because of an intake 100 mg/d, then supplementation with any vitamin C dose cannot further reduce oxidative damage, a fact that likely explains some of the discrepant results of the biomarker studies discussed. More attention to this critical issue and more detailed biomarker studies on vitamin C intakes near tissue-saturating concentrations are warranted.

Are there other mechanisms by which vitamin C could affect chronic disease incidence, mortality, or both?
As discussed above, vitamin C has many functions in the body, such as acting as a cosubstrate for several biosynthetic enzymes (8). Therefore, vitamin C may affect chronic disease incidence, mortality, or both by mechanisms that may not be directly related to its role as an antioxidant.

Cardiovascular disease
Hypercholesterolemia is a significant risk factor for cardiovascular disease (43, 119). The relation between vitamin C supplementation, or plasma vitamin C concentrations, and total serum cholesterol has been investigated in several studies (82, 119–123). In one supplementation study, consumption of 1000 mg vitamin C/d for 4 wk resulted in a reduction in total serum cholesterol (120), whereas in another study, supplementation with 60 or 6000 mg/d for 2 wk had no effect (82). Two observational studies found an inverse correlation between vitamin C status and total serum cholesterol concentrations (122, 123). The mechanism for the possible modulating effect of vitamin C on serum cholesterol concentrations is not entirely certain. One putative pathway is through vitamin C's role as a cofactor for cholesterol 7-monooxygenase, an enzyme involved in the in vivo hydroxylation of cholesterol to form bile acids (8). Vitamin C may also modulate the activity of hydroxymethylglutaryl-CoA reductase, the rate-limiting enzyme in the biosynthesis of cholesterol (43).

The plasma lipoprotein profile is also an important consideration for cardiovascular disease, with decreased concentrations of HDL and increased concentrations of LDL being significant risk factors (43, 119). Numerous observational studies have found a significant association between elevated plasma vitamin C concentrations and increased concentrations of HDL cholesterol and reduced concentrations of LDL cholesterol (119, 121–125). Findings indicated that with every 30-µmol/L increase in plasma vitamin C, HDL was elevated by 4–10% and LDL was reduced by 4% (125). Similar modulatory effects were reported after supplementation with 1000 mg vitamin C/d for 4 wk (120). Ness et al (121) also found an inverse correlation between vitamin C status and triacylglycerol concentrations. Vitamin C may modulate the activity of lipoprotein lipase (43), although the mechanism is unknown.

The thrombotic risk of cardiovascular disease is associated with increased concentrations of the coagulation factor fibrinogen (126). Two studies found an inverse association between serum vitamin C concentrations and coagulation factors as well as a positive association between low serum vitamin C and elevated fibrinogen and coagulation activation markers (126, 127). Two early studies indicated that supplementation of heart disease patients with 2000–3000 mg vitamin C/d for 1–6 wk increased fibrino-lytic activity and reduced platelet adhesiveness (128, 129). In a more recent study (130), healthy volunteers were supplemented with 250 mg vitamin C/d for 8 wk and a nonsignificant decrease in platelet aggregation and an increased sensitivity to PGE₁ were reported. In vitro studies showed that physiologic concentrations of vitamin C may increase PGE₁ and PGI₁ (prostacyclin) production, resulting in a reduction in platelet aggregation and thrombus formation (31), although whether this mechanism is relevant in vivo has yet to be established. Low concentrations of vitamin C are also associated with increased concentrations of plasminogen activator inhibitor 1, a protein that inhibits fibrinolysis (131).

Adhesion of leukocytes to the endothelium is an important initiating step in atherogenesis (31). Smokers have lower plasma vitamin C concentrations than do nonsmokers (88), and monocytes isolated from smokers exhibit increased adhesion to endothelial cells (132, 133). Supplementation of smokers with 2000 mg vitamin C/d for 10 d elevated plasma vitamin C concentrations from 48 to 83 µmol/L and significantly reduced monocyte adhesion to endothelial cells (132). In another study, however, supplementation of smokers with 2000 mg vitamin C 2 h before serum was collected had no significant effect on ex vivo monocyte or endothelial cell adhesion, despite an increase in vita- min C concentrations from 34 to 115 µmol/L (133). Interestingly, supplementation with 7000 mg L-arginine, the physiologic substrate for nitric oxide (NO) synthase, significantly reduced monocyte and endothelial cell adhesion, suggesting an important role for NO. Several in vivo animal studies also suggested an important role for vitamin C in modulating leukocyte and endothelial cell interactions in hamsters exposed to cigarette smoke (134, 135) or injected with oxidized LDL (136).

Impaired vascular function and relaxation are highly relevant to the clinical expression of atherosclerosis, ie, angina pectoris, myocardial infarction, and stroke. Hypertension is a recognized risk factor for cardiovascular disease (43) and low concentrations of plasma vitamin C have been associated with hypertension (122, 125, 137–139). Several studies reported positive effects of high doses of vitamin C, administered either orally or by intraarterial infusion, on vasodilation (Table 3). Four studies (86, 141, 146, 150) investigated vasodilation in patients with cardiovascular disease and found increases of 45–220% in vasodilation after administration of vitamin C (1000–2000 mg oral or 25 mg/min infusion). One of these studies found a 128% increase in brachial artery dilation in coronary artery disease patients, but a nonsignificant 27% increase in chronic heart failure patients (150). Kugiyama et al (148) observed a 100% reversal of epicardial artery vasoconstriction in coronary spastic angina patients infused with 10 mg vitamin C/min. Other studies investigated patients with hypercholesterolemia (144) or hypertension (145, 149), both of which are important risk factors for cardiovascular disease. Infusion of 3000 mg (145) or 24 mg/min (144, 149) vitamin C in these patients resulted in increased blood flow and decreased vasoconstriction. Heitzer et al (142) and Motoyama et al (143) observed increased vasodilation in smokers given infusions of 10–18 mg vitamin C/min. Motoyama et al showed a significant positive correlation (r = 0.597, P = 0.0001) between serum concentrations of vitamin C and the percentage increase in the brachial arterial diameter of smokers and nonsmokers. Similarly, patients with type 2 and type 1 diabetes had increased blood flow after infusion of 24 mg vitamin C/min (140, 147). Finally, healthy individuals given an oral dose of 1000 mg vitamin C in combination with 800 IU vitamin E had increased vasodilation several hours after a single high-fat meal (151).

Cancer
Vitamin C may protect against cancer through several mechanisms in addition to inhibition of DNA oxidation. One potential mechanism is chemoprotection against mutagenic compounds such as nitrosamines (153, 154). N-Nitroso compounds are formed by reaction of nitrite or nitrate (common in cured food and cigarette smoke) with amines and amides (153). Nitrosating compounds can also be formed from NO generated by inflammatory cells expressing inducible NO synthase (116, 153, 155, 156). N-Nitroso compounds undergo activation by cytochrome P450–dependent enzymes and have been implicated in gastric and lung cancer (153). Epidemiologic studies have shown an inverse association between vitamin C intake, mainly from fruit and vegetables, and cancers at these sites (157, 158); additionally, vitamin C reduces in vivo nitrosation by scavenging nitrite and hence preventing its reaction with amines to form nitrosamines (153, 154). Concentrations of fecapentaenes, fecal mutagens that have been implicated in colon cancer (155), are also reduced by vitamin C (159).

In addition, vitamin C may reduce carcinogenesis through stimulation of the immune system. Two of the major functions of the immune system are to fight off infections and to prevent cancer (3). It is hypothesized that the immune system recognizes tumor-forming cells as nonself. Cytotoxic T lymphocytes, macrophages, and natural killer cells can lyse tumor cells (3). Free radicals and oxidative products secreted by immune cells can also lyse tumor cells. Vitamin C can protect host cells against harmful oxidants released into the extracellular medium. Therefore, an optimal immune response requires a balance between free radical generation and antioxidant protection.

Vitamin C is taken up by phagocytes and lymphocytes to concentrations up to 100-fold greater than in plasma, and intracellular concentrations of vitamin C are reduced when phagocytes are activated (3). Many studies have investigated the effects of vitamin C on leukocyte function; however, the data are inconsistent and conflicting (160). Vitamin C may modulate the functions of phagocytes, such as chemotaxis (161–164), as well as the activity of natural killer cells and the functions and proliferation of lymphocytes (160, 165, 166). Vitamin C may also affect the production of immune proteins such as cytokines and antibodies as well as complement components (160, 167, 168). An important measure of overall immune function is the delayed-type hypersensitivity response, which may be modulated by antioxidant micronutrients such as vitamin C (3, 169). Jacob et al (159) showed that vitamin C deficiency in men ingesting 5–20 mg vitamin C/d for 32 d significantly reduced delayed-type hypersensitivity responses, but resulted in no significant change in lymphocyte proliferation. Delayed-type hypersensitivity responses did not return to baseline, even after supplementation with 250 mg vitamin C/d for 4 wk.

Does vitamin C lower chronic disease incidence, mortality, or both?
Many epidemiologic studies and a limited number of clinical trials have indicated that dietary intake of, or supplementation with, antioxidant vitamins is associated with a reduction in the incidence of chronic disease morbidity and mortality (5, 14, 15). Because vitamin C acts as an antioxidant and can ameliorate oxidative damage to lipids, DNA, and proteins, the association of vitamin C with cardiovascular disease, cancer, and cataract, respectively, is of substantial interest. Much of the evidence discussed in the subsequent sections is derived from epidemiologic studies. As such, observed associations between vitamin C intake or plasma concentrations and disease risk are not proof of cause-effect relations; the observed differences may in part reflect differences in dietary behavior or lifestyle patterns. There also may be confounding of interpretations by unmeasured risk factors or imperfect statistical corrections. Furthermore, most of the epidemiologic data are based on dietary vitamin C intake, mainly from fruit and vegetables, and it is difficult to discern whether an observed inverse association with disease incidence is due to vitamin C itself, vitamin C together with other substances in fruit and vegetables, or these other substances themselves, with vitamin C as a surrogate marker. Finally, estimation of vitamin C intake from food-frequency questionnaires has limited accuracy and measurement of vitamin C plasma concentrations, although more accurate than estimation of dietary intake, also has pitfalls and is dependent on proper handling and storage of the samples. These limitations of epidemiologic studies based on dietary intake and plasma concentrations of vitamin C have been discussed previously (158, 170).

Cardiovascular disease
Coronary artery disease and stroke are the leading causes of morbidity and mortality in the United States and other westernized populations. Cardiovascular disease is responsible for nearly one million deaths every year in the United States alone, at the cost of >$15 billion in health care and lost productivity (34). Major risk factors associated with cardiovascular disease are age, male sex, smoking, hypercholesterolemia, hypertension, family history, obesity, and physical inactivity (43, 119). Many epidemiologic studies have shown inverse associations between antioxidant intake, particularly vitamin E, and cardiovascular disease (171). Over the past 15 y, several prospective cohort studies have been published on the association between vitamin C intake and the risk of cardiovascular disease (Table 4). Some of these were reviewed previously by Enstrom (14) and Gey (15), as well as by others (5, 43, 171). Because the purpose of this review is to propose an RDA for vitamin C based on chronic disease incidence, only studies that stated actual amounts of vitamin C intake are considered further here.

Numerous studies reported a reduced risk of cardiovascular disease with vitamin C intakes considerably higher than those in the above-mentioned studies (173–176, 181, 185, 187). Enstrom et al (173) showed a risk reduction in cardiovascular disease of 42% in men and 25% in women consuming >50 mg vitamin C/d from the diet plus regular supplements, corresponding to 300 mg total vita- min C/d (174). An earlier study by Enstrom et al (172) indicated that intakes of vitamin C >250 mg/d were not associated with an additional risk reduction for cardiovascular disease, although subsequent reanalysis of the data indicated that intakes >750 mg/d were associated with a reduction in overall mortality (173). Sahyoun et al (185) reported a significant 62% lower risk of cardiovascular disease in a population of elderly men and women consuming >388 mg vitamin C/d than in those consuming <90 mg/d. Similar intakes were reported by Manson et al (175, 176, 187) in a cohort of female nurses, although only a moderate risk reduction of 24% was observed for stroke, whereas the risk reduction for coronary artery disease was nonsignificant. Finally, Kritchevsky et al (181) measured carotid artery wall thickness as a measure of atherosclerosis and found significantly decreased intima thickness in men and women aged >55 y consuming, respectively, >982 or > 728 mg vitamin C/d than in those consuming <56 or < 64 mg/d.

Interestingly, several epidemiologic studies indicated no association between vitamin C intake or supplementation and risk of cardiovascular disease (177, 183, 184, 186). Losconczy et al (184) and Kushi et al (183) observed no effect on coronary artery disease risk with regular vitamin C supplementation. Kushi et al (183) and Rimm et al (177), in 2 large epidemiologic studies, also reported no additional reduction in risk of coronary artery disease with vitamin C intakes of 200 and 400 mg/d, respectively, compared with intakes of 90 mg/d. One intervention trial found no reduction in risk of stroke or hypertension in a population of Chinese men and women supplemented with 180 mg vitamin C/d and 30 µg Mo/d for 5 years (186).

The study by Levine et al (117), which found that tissue saturation in healthy men occurred at vitamin C intakes of 100 mg/d, may explain why several of the above-mentioned studies showed no protective effect of dietary vitamin C intakes >90 mg/d (177, 179, 183) or vitamin C supplementation (183, 184, 186). Because an intake of 90 mg vitamin C/d results in near tissue saturation, increasing vitamin C intake over this amount may have only a small or no additional effect on tissue concentrations and hence disease risk. Thus, the totality of evidence from prospective cohort studies to date suggests that there is only a minimal intake requirement for vitamin C to optimally reduce the risk of cardiovascular disease, and that there is little or no additional benefit from vitamin C intakes >90–100 mg/d, likely because of tissue saturation at this level (117).

Several investigators studying cardiovascular disease have measured plasma concentrations of vitamin C (Table 5), which is a considerably more accurate and reliable measure of body vitamin C status than dietary intake estimated from questionnaires. One prospective cohort study indicated that plasma vitamin C concentrations >23 µmol/L are associated with moderate, statistically nonsignificant reductions of 20% and 22%, respectively, in the risk of coronary artery disease and stroke (189, 190). Similarly, Gale et al (180) observed a statistically significant 30% lower risk of death from stroke in subjects with plasma vitamin C concentrations >28 µmol/L than in those with concentrations <12 µmol/L; however, no association was observed with coronary artery disease risk. Larger risk reductions of 39–60% were observed for coronary artery disease, myocardial infarction, and angina pectoris with vitamin C concentrations >11–57 µmol/L (188, 191, 193). Furthermore, patients with these conditions were found to have significantly lower plasma vitamin C concentrations than control subjects or survivors (188, 191, 192, 194). Interestingly, in the studies reporting an inverse association between plasma vitamin C concentrations and angina pectoris (188) and coronary artery disease (191), the association was substantially reduced after adjustment for smoking. This finding is to be expected given the known effect of smoking on plasma vitamin C concentrations (196) and suggests that smoking may increase cardiovascular disease risk in part by lowering vitamin C concentrations.

Cancer
More than half a million deaths occur annually from cancer in the United States (197). Lung cancer resulting from smoking causes 30% of all US cancer deaths; colon-rectum, breast, and prostate cancers account for another 25% of deaths (197). The major risk factors for cancer are smoking, chronic inflammation, and an unbalanced diet. A multitude of epidemiologic studies have shown that increased consumption of fresh fruit and vegetables is associated with a reduced risk of most types of cancer (157, 158). Fruit and vegetables contain many constituents that may contribute to protection against cancer, including antioxidant vitamins. Over the years, numerous case-control studies have been carried out to investigate the role of vitamin C in cancer prevention; these have been reviewed by Block (157) and Fontham (158). Most case-control studies showed a consistent inverse association between vitamin C intake and cancers of the oral cavity, larynx-pharynx, and esophagus, as well as of the lung, stomach, and colon-rectum. Of the hormone-dependent cancers, only breast cancer was inversely associated with vitamin C intake, in contrast with ovary and prostate cancers. Although interesting and consistent in their results, these case-control studies may be intrinsically biased because of their retrospective design (158). Therefore, only prospective cohort studies, some of which were reviewed by Enstrom (14) and Gey (15), that also stated actual dietary intakes are considered further in this section (Table 6).

Interestingly, virtually all of the studies in which vitamin C intakes were >87 mg/d in the lowest intake group (quantile) found no or nonsignificant effects on cancer risk reduction with higher intakes of vitamin C (172, 185, 203, 204, 206). Only one study found a significant protective effect: Shibata et al (201) observed a moderate 24% lower risk of all cancers in women consuming >225 mg vitamin C/d than in those consuming <155 mg/d, although no effect was observed in men. In good agreement with these studies, vitamin C supplementation was not associated or nonsignificantly associated with reduced cancer risk in several prospective cohort studies (184, 201, 203, 206), although Bostick et al (204) observed a 33% reduction in colon cancer risk in a large population of women consuming >60 mg supplemental vita- min C/d. No association was apparent when supplements were taken for >10 y by patients with breast cancer (203). The Linxian trial found no significant effect of supplementing a population of Chinese men and women with 120 mg vitamin C/d and 30 µg Mo/d for 5 y on the risk of cancers of the esophagus or stomach (205). The results of clinical trials, however, depend on the use of sufficient doses, sufficient durations, and low baseline concentrations of vitamin C.

In summary, in most of the studies in Table 6 that reported no significant reduction in cancer risk, intakes of vitamin C in the lowest quantile were >86 mg/d; those studies that reported significant risk reductions (173, 182, 199, 207) found this effect in individuals with vitamin C intakes 80–110 mg/d. As discussed above, this intake range of 80–110 mg/d is associated with vitamin C tissue saturation in healthy men (117). With one exception (204), studies investigating consumption of supplemental vitamin C, including the Linxian trial (205), did not show a protective effect against cancer, possibly because the dietary intake of vitamin C was already sufficient for tissue saturation. Therefore, the consensus protective intake emerging from the studies in Table 6 appears to be 80–110 mg vitamin C/d. More studies investigating cancer risk in persons with low vitamin C intakes are warranted.

Several case-control and prospective cohort studies have investigated the association between plasma concentrations of vitamin C and cancer risk (Table 7). Four studies found significantly higher concentrations of vitamin C in control subjects or survivors than in patients with cancer (208, 209, 211, 212), who in all 4 studies had plasma vitamin C concentrations <45 µmol/L. Two recent studies also found lower plasma vitamin C concentrations in cancer patients than in control subjects, but the differences were nonsignificant (46, 213). Because of the retrospective nature of case-control studies, it is not clear whether decreased plasma vitamin C concentrations in cancer patients are a cause or a consequence of the disease. In a prospective cohort study, Sahyoun et al (185) found a 32% lower risk of cancer in elderly persons with concentrations of vitamin C >89 µmol/L than in those with concentrations <52 µmol/L. The Basel study investigated cancers at several sites over 7–17 y of follow-up in a population of 2974 Swiss men (210, 214, 215). Vitamin C concentrations >23 µmol/L were associated with a nonsignificant reduction in risk of cancer at all sites after 17 y (210). Nonsignificant reductions of 42% and 45% in colon cancer and lung cancer risk, respectively, were observed. No associations were observed for risk of stomach cancer or prostate cancer, the latter being a hormone-dependent cancer and hence less likely to be affected by diet, including vitamin C (157). In a comprehensive review article, Gey (15) suggested that plasma vitamin C concentrations >50 µmol/L are associated with protection against cancer, which is consistent with the findings of the studies listed in Table 7 and an intake of 100 mg/d (117).

Several epidemiologic studies have investigated the association of vitamin C intake with the incidence of cataract (Table 8). Two case-control studies indicated a strong inverse association between high intakes of vitamin C and cataract (217, 218). Robertson et al (217) found that intakes of >300 mg vitamin C/d were associated with a 70% reduced risk of cataract. Similarly, Jacques and Chylack (218) found that daily intakes of >490 mg were associated with a 75% lower risk of cataract than intakes <125 mg/d. These investigators also measured plasma concentrations of vitamin C: concentrations >90 µmol/L were associated with a 71% lower risk than concentrations <40 µmol/L. In contrast, Vitale et al (220) found no protective effect of 260 mg vitamin C/d over that of 115 mg/d and no association with plasma concentrations >80 µmol/L compared with <60 µmol/L.

As mentioned earlier, epidemiologic studies can provide only a semiquantitative estimation of vitamin C intake depending on the accuracy of dietary recall questionnaires; additionally, in many of the above-mentioned studies, intake at only one time point was assessed. Determination of plasma concentrations of vitamin C may more accurately reflect body stores of the vitamin. The protective concentrati