CVD is believed to become the biggest cause of morbidity and mortality in men and women in the world in 2020(Reference Lopez and Murray1), emphasising the great need for retarding the increase in disease incidence. Individuals with a high dietary intake of fruit and vegetables have a clear reduction in the incidence of CHD(Reference He, Nowson and Lucas2–Reference Liu, Lee and Ajani5), stroke(Reference He, Nowson and MacGregor6–Reference Vollset and Bjelke9) and cardiovascular mortality(Reference Gaziano, Manson and Branch10, Reference Liu, Manson and Lee11). Reactive oxygen species and free radicals have been implicated in the pathophysiology of CVD(Reference Cross, Halliwell and Borish12), with vitamins E and C and β-carotene being hypothesised as the fundamental protective components in fruit and vegetables. It has also been hypothesised that flavonoids and fibre are also likely to be potential fundamental protective components in fruit and vegetables.
The body is equipped with antioxidative enzymes, such as glutathione peroxidase and superoxide dismutase, and vitamins including vitamins E and C and β-carotene which cooperate and in some cases act synergistically to provide protection against oxidative stress. Atherosclerosis is the underlying cause of CVD, involving the accumulation of modified LDL in the intima of the arterial wall(Reference Finking and Hanke13) enabling plaque progression(Reference Steinberg, Parthasarathy and Carew14) and the occurrence of cardiovascular events(Reference Berliner and Heinecke15). LDL particles contain about 2700 fatty acids of which approximately half are polyunsaturated and are susceptible to oxidation(Reference Jialal and Devaraj16).
The identification of the oxidative modification hypothesis of LDL(Reference Steinberg, Parthasarathy and Carew14) and the strong correlation between the levels of oxidised LDL(Reference Stringer, Gorog and Freeman17) and the ex vivo oxidative susceptibility of LDL(Reference Regnstrom, Nilsson and Tornvall18) to the apparent extent of atherosclerosis provide a rationale for a role of oxidative stress in atherosclerosis. Oxidised LDL acts as a chemokine that stimulates the recruitment of circulating monocytes into the intimal space(Reference Quinn, Parthasarthy and Fong19, Reference Quinn, Parthasarthy and Steinberg20) and inhibits the exit of resident macrophages(Reference Quinn, Parthasarthy and Steinberg20), enabling foam cell formation and cell-mediated LDL peroxidation. Oxidised LDL is cytotoxic(Reference Cathcart, Morel and Chisolm21–Reference Morel, Hessler and Chisolm23) and also reduces NO bioavailability(Reference Cai and Harrison24–Reference Galle, Mulsch and Busse26), which results in endothelial dysfunction. In accordance with the response-to-injury hypothesis of atherogenesis, this results in the progression of the atherosclerotic lesion(Reference Steinberg, Parthasarathy and Carew14) and consequent cardiovascular events(Reference Berliner and Heinecke15). The role of these vitamins (vitamins E and C and β-carotene) is emphasised by their inhibitory action on the oxidative modification of LDL(Reference Diaz, Frei and Vita27, Reference Steinbrecher, Parthasarathy and Leake28) and their improvement of endothelial dysfunction(Reference Berliner and Heinecke15, Reference Taddei, Virdis and Ghiadoni29). Their therapeutic role has been supported by animal studies(Reference Crawford, Kirk and Rosenfeld30–Reference Smith and Kummerow33) and has been further supported by changes in lipid peroxide levels(Reference Mezzetti, Zuliani and Romano34), ex vivo oxidisability of LDL(Reference Haidari, Javadi and Kadkhodaee35, Reference Chiu, Jeng and Shieh36), and plasma levels of these vitamins(Reference Luoma, Nayha and Sikkila37–Reference Esterbauer, Rotheneder and Striegl41) being potential good predictors of future cardiac events and cardiovascular mortality.
Susceptibility to LDL peroxidation is dependent on the levels of these vitamins(Reference Esterbauer, Rotheneder and Striegl41) and only once they are fully depleted is rapid oxidation possible(Reference Jessup, Rankin and De Whalley42). As a consequence these vitamins are frequently referred to, perhaps simplistically, as antioxidant vitamins. The role of these vitamins in reducing LDL oxidation has been most consistently shown for vitamin E while the data has been mixed for vitamin C and β-carotene. Therefore greater emphasis has been put on vitamin E in exploring a preventive or therapeutic role for these vitamins in CVD.
The role of these vitamins in CVD has long been emphasised mainly on the basis of their hypothesised antioxidant properties and the majority of trials have been initiated on this basis. However, in recent years research has greatly expanded in this area and studies now strongly support that the concept that these vitamins have other fundamental non-antioxidative properties, including actions on different aspects of the inflammatory responses that are involved in the pathogenesis of CVD. Each vitamin may have its own non-antioxidative properties, which will be discussed more in detail below, and as a consequence these vitamins target different aspects of the pathogenesis of CVD and as a result more emphasis can be put on the role of combination therapy. Through these properties a further novel role for these vitamins in CVD may be proposed. In the present review we will focus on putative antioxidant roles of these vitamins, but it should be emphasised that their non-antioxidative properties may be relevant to the modulation of CVD risk.
Vitamin E
Vitamin E is the main chain-breaking lipid-soluble vitamin in plasma and LDL(Reference Burton, Joyce and Ingold43), present in a complex of four isomers (α-tocopherol, γ-tocopherol, β-tocopherol and δ-tocopherol), of which α-tocopherol is biologically the most active(Reference Traber44). Supplementation with pharmacological doses ( ≥ 150 IU/d, ≤ 1200 IU/d) of vitamin E has been shown to reduce LDL peroxidation(Reference Princen, van Poppel and Vogelezang45, Reference Dieber-Rotheneder, Puhl and Waeg46) (1 mg vitamin E is equivalent to 1·49 IU vitamin E).
Atherosclerosis is now accepted to be a chronic inflammatory disease(Reference Diaz, Frei and Vita27, Reference Ross47) and vitamin E has shown to mediate anti-inflammatory effects beyond its antioxidative properties(Reference Islam, Devaraj and Jialal48–Reference Zapolska-Downar, Zapolski-Downar and Markiewski51). Through these non-antioxidative properties vitamin E may target aspects of atherosclerosis beyond the oxidation of LDL, therefore extending its potential protective role in CVD. Vitamin E potentially reduces foam cell formation by decreasing monocyte recruitment(Reference Devaraj, Li and Jialal49, Reference Wu, Koga and Martin52), through reducing chemokine secretion(Reference Devaraj and Jialal53) and by reducing the expression of scavenger receptors on macrophages (CD36)(Reference Munteanu, Taddei and Tamburini54, Reference Devaraj, Hugou and Jialal55). Vitamin E can also potentially reduce the progression of atherosclerosis by reducing adhesion molecule expression(Reference Islam, Devaraj and Jialal48, Reference Zapolska-Downar, Zapolski-Downar and Markiewski51), inhibiting smooth muscle cell proliferation(Reference Keaney, Simon and Freedman56, Reference Ozer, Palozza and Boscoboinik57) and platelet aggregation(Reference Freedman, Farhat and Loscalzo58, Reference Steiner59) and by enhancing NO bioavailability(Reference Keaney, Gaziano and Xu60). These effects have been shown to be partly mediated via non-antioxidant mechanisms causing inhibition of signalling pathways, particularly protein kinase C(Reference Murohara, Ikeda and Katoh61, Reference Boscoboinik, Szewczyk and Hensey62), that have potentially been activated by oxidised LDL. Vitamin E has been shown to prevent oxidised LDL-induced NF-κB activation through suppressing protein kinase C(Reference Sugiyama, Kugiyama and Ogata63) and inhibiting IκB degradation(Reference Li, Saldeen and Mehta64), further reducing the inflammatory response that is mediated in CVD.
Another anti-atherogenic property of vitamin E is its ability to modulate gene expression, such as up-regulating endothelial NO synthase mRNA expression and consequently NO levels(Reference Goya, Sumitani and Otsuki65), hence protecting the endothelium. Vitamin E has been shown to prevent endothelial dysfunction through protecting the endothelium against reactive oxygen species and oxidised LDL(Reference Keaney, Guo and Cunningham66) and through stimulating endothelial cell proliferation(Reference Ulrich-Merzenich, Metzner and Schiermeyer67, Reference Kuzuya, Naito and Funaki68) and reducing endothelial apoptosis(Reference Uemura, Manabe and Yoshida69, Reference Li, Saldeen and Romeo70). These effects are mediated by mechanisms beyond that of inhibition of oxidation of LDL, which include inhibition of oxidised LDL-induced protein kinase C stimulation(Reference Keaney, Guo and Cunningham66), possibly via an activation of a phosphatase PP2A(Reference Azzi, Aratri and Boscoboinik71), modulation of the Bcl-2 family of apoptosis-related proteins(Reference Haendeler, Zeiher and Dimmeler72), by inhibiting caspase-3 activity(Reference Uemura, Manabe and Yoshida69) and by inhibiting the oxidised LDL-induced up-regulation of angiotensin II receptor (AT1R) mRNA and protein. These properties have been further supported by animal studies(Reference Koga, Kwan and Zubik73).
These effects of α-tocopherol have only been confirmed by in vitro studies and animal studies but not yet in vivo. The importance of vitamin E in protecting against atherosclerosis has been further supported by the vitamin E-deficient mouse model, which suffered from increased levels of oxidative stress and atherosclerosis(Reference Terasawa, Ladha and Leonard74).
Vitamin C
The independent role of vitamin C in CVD has not been extensively assessed in clinical trials. However, as LDL oxidation occurs substantially in the sub-endothelial space(Reference Niki, Yamamoto and Komuro75), vitamin C may be most important in maintaining the reduced state of vitamin E. Water-soluble antioxidant vitamins, predominantly vitamin C, work to prevent the consumption of hydrophobic antioxidant vitamins such as vitamin E and β-carotene(Reference Jialal and Grundy76) and ensure their recycling(Reference Kagan, Serbinova and Forte77), therefore playing an important role in maintaining antioxidative protection. Therefore vitamin C can act synergistically with these other vitamins, enhancing the benefit achieved with supplementation. Like vitamin E, vitamin C has been shown to have additional non-antioxidant properties. Vitamin C has been shown in vivo to suppress endothelial apoptosis mediated by inflammatory cytokines and oxidised LDL(Reference Rossig, Hoffmann and Hugel78) and it has been shown to promote the proliferation of endothelial cells and the inhibition of vascular smooth muscle growth(Reference Ulrich-Merzenich, Metzner and Schiermeyer67)via the extracellular signal-regulated kinase-signalling pathway(Reference Ulrich-Merzenich, Zeitler and Panek79). It has also been suggested that vitamin C has a role in preventing restenosis postangioplasty(Reference Tomoda, Yoshitake and Morimoto80). In fact the combination of vitamins C and E exhibited a stronger positive effect than vitamin C or vitamin E did on their own. Vitamin C has the ability to modulate gene expression and through down-regulating intercellular adhesion molecule-1 gene expression it can reduce monocyte adherence to the endothelium(Reference Rayment, Shaw and Woollard81). Vitamin C has also been shown to enhance NO synthesis in endothelial cells(Reference Heller, Munscher-Paulig and Grabner82) and in vivo it has been shown to have sustained beneficial effects on endothelial-derived NO-dependent flow-mediated dilation(Reference Gokce, Keaney and Frei83). Vitamin C supplementation has also been shown to reduce vascular smooth muscle cell apoptosis and therefore prevent plaque instability in late-stage atherosclerosis(Reference Siow, Richards and Pedley84).
β-Carotene
β-Carotene is a fat-soluble vitamin present together with vitamin E in the lipid core of LDL particles(Reference Esterbauer, Rotheneder and Striegl41). It is an excellent trapper of singlet oxygen and potentially a second-line antioxidative defence for LDL particles once vitamin E has been utilised(Reference Jessup, Rankin and De Whalley42). The role of carotenoids in oxidative protection has been inconsistent, data indicating neutral(Reference Gaziano, Manson and Branch10, Reference Princen, van Poppel and Vogelezang45, Reference Reaven, Khouw and Beltz85), anti-(Reference Tsuchihashi, Kigoshi and Iwatsuki86, Reference Jialal, Norkus and Cristol87) and pro-oxidant(Reference Rautalahti, Albanes and Virtamo88, Reference Palozza, Calviello and Bartoli89) properties. The pro-oxidant effects have been proposed to be due to the tendency of β-carotene radicals reacting with oxygen to give rise to peroxyl radicals that mediate lipid peroxidation(Reference Tsuchihashi, Kigoshi and Iwatsuki86). Serum carotenoid levels have been inversely associated with atherogenic factors(Reference Hozawa, Jacobs and Steffes90), risk of atherosclerosis(Reference D'Odorico, Martines and Kiechl91) and cardiovascular mortality(Reference Ito, Kurata and Suzuki92); however, these studies looked at the possible effect of a combination of carotenoids and did not assess the independent effect of β-carotene.
High dietary intake of vitamin E(Reference Knekt, Ritz and Pereira93–Reference Stampfer, Hennekens and Manson97), vitamin C(Reference Osganian, Stampfer and Rimm98, Reference Enstrom, Kanim and Klein99) and β-carotene(Reference Rimm, Stampfer and Ascherio96, Reference Osganian, Stampfer and Rimm100, Reference Klipstein-Grobusch, Geleijnse and den Breeijen101) has been inversely associated with the incidence of CHD. High dietary intake of β-carotene has been associated with a reduced CVD mortality(Reference Buijsse, Feskens and Kwape102) and all-cause mortality(Reference Buijsse, Feskens and Schlettwein-Gsell103); however, this was restricted to elderly individuals.
The favourable safety profile of these vitamins(Reference Hathcock, Azzi and Blumberg104, Reference Kappus and Diplock105) has allowed several clinical trials to be conducted attempting to confirm their role. At this point the results have been inconsistent, with a few small trials suggesting a protective role while large-scale trials have concluded no benefit with vitamin supplementation in patients at high risk of CVD(Reference Wright, Lawson and Weinstein106–Reference Blot, Li and Taylor116), or with pre-existing CVD(Reference Cook, Albert and Gaziano117–Reference Brown, Zhao and Chait120).
There have been several explanations for this lack of correlation between observational studies and randomised controlled trials. The lack of benefit in randomised controlled trials could suggest that these vitamins are not the protective components in fruit and vegetables. As the results of observational studies can be as a consequence of confounding factors it is possible that other components of fruit and vegetables are the mediators of cardiovascular protection, such as flavonoids, fibre, etc.
However, the lack of benefit could also be a consequence of the differences in duration, vitamin dosages and target population between observational studies and randomised controlled trials.
Observational studies have been conducted on an average for 11 years while randomised controlled trials have continued for an average of 4 years, which can suggest that supplementation needs to be conducted for a longer period of time to gain benefit. Steinberg(Reference Steinberg121) hypothesised that antioxidants were targeting early stages of atherosclerosis so that the average 4·5-year duration of the majority of trials was too short to achieve beneficial effects. However, none of the pre-existing trials have indicated any trend towards a protective role and the two trials conducted over more than 10 years(Reference Lee, Cook and Gaziano108, Reference Hennekens, Buring and Manson113) have further disputed the role of antioxidants in CVD. Therefore before the trial duration is extended other areas should be addressed. The lack of detailed knowledge of the mechanism of oxidative modification has restricted us in defining an optimal antioxidant vitamin. The lack of efficient biomarkers for oxidative stress has not allowed us to assess in vivo effectiveness of these vitamins' antioxidant properties and define the optimal vitamin dosage. Whether the dosage of these vitamins plays a role in their beneficial effects is addressed in the present review.
The targeted population is still undefined; however, pre-existing evidence is suggestive of targeting subgroups such as smokers, diabetics and the elderly.
These vitamins have been shown to mediate effects beyond their antioxidative properties; however, at this point these have only been shown in vitro and not yet explored in in vivo studies. The present review will address the hypotheses that have been put forward to try to explain the lack of benefit with these vitamins in randomised controlled trials, provide further evidence regarding their role in CVD and explore what the future may entail for vitamin therapy in CVD.
Dosage, oxidative markers and isomers
The optimal vitamin dosage has not yet been defined. Nutritional doses of vitamin E (about 4–8 mg/d)(Reference Knekt, Ritz and Pereira93–Reference Rimm, Stampfer and Ascherio96) and vitamin E supplementation for at least 2 years with >100 IU/d(Reference Rimm, Stampfer and Ascherio96, Reference Stampfer, Hennekens and Manson97) with a maximum dose of 1000 mg/d(Reference Hathcock, Azzi and Blumberg104) have been beneficial in CHD. However, the majority of observational studies have shown disappointing findings in regards to supplemental intake of vitamin E ( ≤ 25 mg/d up to ≥ 250 mg/d)(Reference Knekt, Ritz and Pereira93, Reference Kushi, Folsom and Prineas94). Randomised clinical trials supplementing with 330–800 IU vitamin E per d have also been disappointing(Reference Lee, Cook and Gaziano108, Reference Hodis, Mack and LaBree110, 115, Reference de Gaetano122–Reference Lonn, Bosch and Yusuf125). The doses of these vitamins used in trials have been questioned, on the one hand for being suboptimal and on the other for being too high. Studies by Jialal et al. (Reference Jialal, Fuller and Huet126) and Simons et al. (Reference Simons, Von Konigsmark and Balasubramaniam127) and findings from observational studies support the concept that the dosages used in trials are not suboptimal. The use of mega-doses of these vitamins has been disputed due to their potential pro-oxidant(Reference Princen, van Poppel and Vogelezang45, Reference Upston, Terentis and Stocker128, Reference Podmore, Griffiths and Herbert129) and pro-atherogenic effects(Reference Keaney, Gaziano and Xu130) and their negative drug interactions(Reference Cheung, Zhao and Chait131, Reference Corrigan132). Even though adverse effects are uncommon and shown to occur at doses well above those used in trials, it is possible that these override their beneficial effects, giving no net gained benefit.
The Vitamin E Atherosclerosis Prevention Study (VEAPS) trial(Reference Hodis, Mack and LaBree110) indicated that a level of oxidative protection is needed to be achieved to gain anti-atherogenic effects, which is suggested by trials to be achieved with 800 IU RRR-α-tocopherol per d(Reference Fang, Kinlay and Beltrame133–Reference Stephens, Parsons and Schofield135). Vitamins' antioxidative effectiveness is assessed ex vivo or through plasma or urinary levels of oxidised biomarkers and it is not clear whether this accurately estimates arterial wall oxidation. These vitamins have been shown to reduce levels of oxidative stress in plasma but not in plaques(Reference Steinberg121). Out of eighteen large-scale trials, only three assessed the effect that vitamin supplementation had on the level of oxidative stress (Table 1) (Reference Hercberg, Galan and Preziosi107, Reference Lee, Cook and Gaziano108, Reference Hodis, Mack and LaBree110–Reference Salonen, Nyyssonen and Salonen112, Reference Omenn, Goodman and Thornquist114–Reference Blot, Li and Taylor116, Reference Waters, Alderman and Hsia118–Reference Brown, Zhao and Chait120, Reference de Gaetano122–Reference Jialal, Fuller and Huet126, Reference Fang, Kinlay and Beltrame133–Reference Manson, Gaziano and Spelsberg136).
Table 1 Trials assessing antioxidant effectiveness*
na, Not been assessed in the trial.
* This Table looks at how intervention studies that are assessing the role of antioxidants in CVD have tried to assess the effectiveness of antioxidants, with some of them measuring antioxidant plasma levels and a few measuring the level of LDL oxidation.
Failure to achieve the oxidative threshold could be the underlying reason behind the disappointing findings of trials. Through identifying more accurate oxidative biomarkers we could assess whether these vitamins mediate their predicted antioxidative effects and identify dose–response curves for optimal oxidative and inflammatory protection.
It has been proposed that the vitamin isomer used in trials is relevant in regards to its effects. The trials that have concluded a positive effect with vitamin E all used RRR-α-tocopherol(Reference Salonen, Nyyssonen and Salonen112, Reference Fang, Kinlay and Beltrame133–Reference Stephens, Parsons and Schofield135) and five out of nine trials that indicated neutral effects used all-rac α-tocopherol(Reference Hodis, Mack and LaBree110, 115, 119, Reference de Gaetano122, 124). Stereoisomers differ structurally and as a result this can restrict their participation in signalling pathways and in other processes, which can result in them not mediating their non-antioxidative actions including the anti-inflammatory effects discussed previously. It is therefore possible that due to the vitamin E isomer that is used for supplementation in trials the non-antioxidative effects of vitamin E are not observed. All-rac α-tocopherol has a lower bioactivity than RRR-α-tocopherol(Reference Hoppe and Krennrich137, Reference Leth and Sondergaard138) and has been shown to lack anti-inflammatory properties(Reference Vega-Lopez, Kaul and Devaraj139) at dosages where this is achieved by RRR-α-tocopherol(Reference Upritchard, Sutherland and Mann140).
The interaction of exogenous and endogenous vitamins: are we using the wrong vitamin?
Traber(Reference Traber141) hypothesised that single supplements may interfere with the uptake, transport, distribution and metabolism of other non-supplemented antioxidant nutrients. The disappointing results of clinical trials could therefore be a result of vitamins' negative interaction with other potentially protective vitamins. Even though studies have emphasised a role for α-tocopherol, γ-tocopherol has been shown to have an anti-atherogenic role(Reference Jiang, Christen and Shigenaga142) and a superior anti-inflammatory effect to that of α-tocopherol(Reference Jiang, Elson-Schwab and Courtemanche143). The main constituent of vitamin E supplementation is usually α-tocopherol, which has been implicated in reducing γ-tocopherol levels(Reference Dieber-Rotheneder, Puhl and Waeg46, Reference Huang and Appel144, Reference Handelman, Machlin and Fitch145) through competing for the same intestinal uptake mechanism(Reference Handelman, Machlin and Fitch145). Therefore the lack of benefits in trials could potentially be due to the gain from one vitamin causing the loss in protection mediated by another vitamin.
As we still lack knowledge regarding the mechanism of LDL oxidation in vivo the optimal vitamin in this context has not been defined. Total carotenoid intake has been associated with a reduced cardiovascular incidence and mortality(Reference Liu, Lee and Ajani5, Reference Gaziano, Manson and Branch10) and the lack of benefit with β-carotene is suggestive that this is the wrong carotenoid. Lycopene is a singlet oxygen scavenger which is part of the carotenoid family and has been predicted to be a stronger antioxidant vitamin than β-carotene(Reference Di Mascio, Kaiser and Sies146). High plasma levels of lycopene have been associated with a reduced risk of atherosclerosis(Reference Rissanen, Voutilainen and Nyyssonen147, Reference McQuillan, Hung and Beilby148) and CVD(Reference Sesso, Buring and Norkus149, Reference Rissanen, Voutilainen and Nyyssonen150). The effects of lycopene have not yet been assessed in large-scale trials.
Is combination therapy superior to single vitamin supplementation? Should we avoid β-carotene?
These vitamins show different efficacy depending on the type of oxidative stress and the body compartment in which it takes place. The lack of knowledge regarding where and how LDL undergoes oxidative modification has restricted us in defining the optimal vitamin type. As these vitamins each possess a specific role in the antioxidant defence system, through the use of a combination of vitamins the overall protection would potentially be broadened.
The potential superiority of combination therapy may be predicted from the following:
(1) Protective effects seen in observational studies with high dietary intake of fruit and vegetables containing several of these vitamins;
(2) Lack of benefit in randomised clinical trials with single compound supplementation;
(3) Pro-oxidant effect of these vitamins in the absence of required cofactors;
(4) Experimental data for the cooperative and synergistic effects of vitamins.
In fruit and vegetables there is a natural interaction between hydrophobic (for example, vitamin E) and hydrophilic (vitamin C) antioxidant vitamins that is lost with single vitamin supplementation and this could account for the lack of benefit. Supplementation with only one of these vitamins could result in an imbalance of endogenous antioxidants which weakens the antioxidant defence system and enables pro-oxidant effects to emerge(Reference Stocker, Bowry and Frei151), as with tocopherol-mediated atherosclerosis seen with high doses of vitamin E(Reference Neuzil, Thomas and Stocker152). Through increasing the dietary intake of fruit and vegetables this results in increased levels of these vitamins in the ‘right environment’. Therefore through combination supplementation using doses of vitamins in physiological ratios we can optimise antioxidant status without resulting in an imbalance in the endogenous antioxidant levels. These vitamins have shown to act synergistically to mediate protection against oxidative stress. Vitamin C has been shown to regenerate vitamin E from its oxidised to its active state(Reference Kagan, Serbinova and Forte77, Reference Buettner153, Reference Packer, Slater and Willson154), to minimise its pro-oxidant effects(Reference Hirano, Kondo and Iwamoto155) and to cause synergistic inhibition of LDL peroxidation(Reference Sato, Niki and Shimasaki156, Reference Rifici and Khachadurian157). β-Carotene has also been shown to act synergistically with vitamin E(Reference Palozza and Krinsky158). It can then be hypothesised that this enhanced protection against oxidative stress should provide a greater anti-atherogenic effect and as a consequence reduce the incidence of clinical endpoints. However, clinical trials using a ‘cocktail’ of vitamins have not indicated any such positive effects(Reference Hercberg, Galan and Preziosi107, Reference Blot, Li and Taylor116–119). Jialal & Grundy(Reference Jialal and Grundy159) and Fuller et al. (Reference Fuller, May and Martin160) concluded that combinations of these vitamins at doses similar to those used in trials (400–800 IU vitamin E per d, 1 g vitamin C per d and 30 mg β-carotene per d) did not provide further oxidative protection of LDL compared with a high dose of α-tocopherol (800 IU/d) on its own. This is therefore suggestive that the combination of these vitamins does not cause a greater reduction in lipid peroxidation than that attributable to single vitamin supplementation.
The majority of the ‘cocktail’ supplementations used in trials have included β-carotene(Reference Hercberg, Galan and Preziosi107, Reference Christen and Gaziano111, Reference Blot, Li and Taylor116, 119, Reference Brown, Zhao and Chait120, Reference Singh, Niaz and Rastogi161) despite the pre-existing data disputing a role for β-carotene as an antioxidant vitamin, and even indicating pro-oxidant effects at the dosages used in trials(Reference Rautalahti, Albanes and Virtamo88, Reference Palozza, Calviello and Bartoli89, Reference Yang and Lowe162). Together with the negative effects seen in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC)(Reference Wright, Lawson and Weinstein106) and the Beta-Carotene and Retinol Efficacy Trial (CARET)(Reference Omenn, Goodman and Thornquist114) its use in CVD is highly questionable, particularly in smokers. The trials including β-carotene supplementation have overall failed to show a beneficial role in CVD (Table 2) (Reference Hercberg, Galan and Preziosi107, Reference Christen and Gaziano111, Reference Blot, Li and Taylor116, 119); however, a combination of vitamins excluding β-carotene has indicated a beneficial role in atherosclerosis (Table 3) (Reference Salonen, Nyyssonen and Salonen112, Reference Arad, Spadaro and Roth163). In vivo the carotenoids do not appear alone but in a heterogeneous mixture, possibly acting synergistically(Reference Kiokias and Gordon164, Reference Stahl, Junghans and de Boer165). Through supplementation with only one carotenoid this can potentially lead to negative effects. Therefore the lack of overall protection with combination therapy could be as a result of a net negative balance between β-carotene pro-oxidant and vitamin E and C antioxidant effects. The beneficial effects in the Transplant-Associated Arteriosclerosis Trial(Reference Fang, Kinlay and Beltrame133) and the Antioxidant Supplementation in Atherosclerosis Prevention (ASAP) trial(Reference Salonen, Nyyssonen and Salonen112) do, however, support a role for vitamins C and E in combination in preventing the progression of atherosclerosis. However, trials have shown contradictory results with the combined supplementation of vitamin C and vitamin E with regard to clinical endpoints, with one suggesting protective effects(Reference Arad, Spadaro and Roth163) and the other indicating neutral effects(Reference Cook, Albert and Gaziano117).
Table 2 Intervention studies: combination of antioxidants including β-carotene(Reference Steinberg121, Reference Stanner, Hughes and Kelly255, Reference Clarke and Armitage292)*†
SU.VI.MAX, Supplémentation en Vitamines et Minéraux Antioxydants; HATS, HDL-Atherosclerosis Treatment Study.
* This Table includes intervention studies (double-blinded randomised controlled trials) that assess the effect of a combination of antioxidants (that include β-carotene) on CVD. It provides information regarding the structure of the intervention studies and their outcomes.
† 1 mg vitamin E per d is equivalent to 1·49 IU vitamin E per d.
Table 3 Intervention studies: combination of antioxidants excluding β-carotene(Reference Steinberg121, Reference Stanner, Hughes and Kelly255, Reference Clarke and Armitage292)*†
ASAP, Atorvastatin Simvastatin Atherosclerosis Progression; IMT, intima-to-media thickness; WACS, Women's Antioxidant Cardiovascular Study; WAVE, Women's Angiographic Vitamin and Estrogen.
* This Table includes intervention studies (double-blinded randomised controlled trials) that assess the effect of a combination of antioxidants (that exclude β-carotene) on CVD. It provides information regarding the structure of the intervention studies and their outcomes.
† 1 mg vitamin E per d is equivalent to 1·49 IU vitamin E per d.
The role of vitamins in the progression and complication of atherosclerosis: should we start these vitamins earlier?
The role of these vitamins in preventing the progression of atherosclerosis and destabilisation of plaques has not been fully confirmed. Oxidised LDL has been shown to stimulate smooth muscle proliferation(Reference Zhao, Seng and Zhang166, Reference Qiao, Zhang and Xia167) and platelet aggregation(Reference Siess, Zangl and Essler168) and to be an independent marker for the destabilisation of plaques(Reference Anselmi, Garbin and Agostoni169). These vitamins have in vitro been shown to reduce platelet aggregation(Reference Ryszawa, Kawczynska-Drozdz and Pryjma170, Reference Jandak, Steiner and Richardson171) and modulate smooth muscle phenotype(Reference Azzi, Aratri and Boscoboinik71), potentially playing a role in retarding the progression of late-stage atherosclerosis, hence attempts to use them in secondary prevention. However, the neutral effects seen in secondary prevention trials may be indicative of the wrong timing of supplementation. Animal studies and observational studies have indicated a therapeutic role through assessing their effect on early lesions while in trials the primary endpoints have been the incidence of major vascular events. Steinberg & Witztum(Reference Steinberg and Witztum172) suggested that antioxidant vitamins are only protective when given before the development of disease, prioritising a role for them in primary prevention.
A meta-analysis of secondary prevention trials concluded that there was a lack of anti-atherogenic effect of vitamin supplementation(Reference Bleys, Miller and Pastor-Barriuso173) and individuals with late-stage atherosclerosis and pre-existing CVD actually had increased cardiac and all-cause mortality with vitamin supplementation(119, Reference Manson, Gaziano and Spelsberg136, Reference Rapola, Virtamo and Ripatti174, Reference Miller, Pastor-Barriuso and Dalal175). This negative effect on fatal and non-fatal CHD is not seen in individuals without pre-existing CHD(Reference Virtamo, Rapola and Ripatti176) and the use of vitamins in these individuals has even suggested a 30 % reduction in overall mortality(Reference Hayden, Welsh-Bohmer and Wengreen177). The negative effects of these vitamins on late-stage atherosclerosis may be due to their limiting effect on ischaemic pre-conditioning(Reference Sun, Tang and Park178), negative interaction with drugs commonly taken by these patients such as nitrates, warfarin and diuretics(Reference Hayden, Welsh-Bohmer and Wengreen177) and their pro-oxidant effects(Reference Bowry and Stocker179) that can destabilise the plaque. These findings are suggestive that vitamin supplementation may have an adverse effect on plaque-related complications and, if so, its use should be restricted to those with early stages of disease, excluding individuals with late-stage atherosclerosis. However, this is difficult in practice. In Western populations atherosclerosis begins early in life, implying that such supplementation should be initiated in childhood and continued for decades. At the same time most adults, certainly those with overt CVD, will have late atherosclerotic lesions.
Directing vitamin use to subgroups
Jialal et al. (Reference Jialal, Freeman and Grundy180) concluded that LDL preparations from different individuals showed different susceptibility or resistance to oxidation. Studies have indicated inter-individual variation in the response seen with antioxidants(Reference Esterbauer, Puhl and Dieber-Rotheneder181), suggesting that individuals exposed to increased levels of oxidative stress or who were antioxidant deficient would gain more benefit(Reference Halliwell182). Vitamin E has been shown to have a variable antioxidant effect that is dependent on the rate of lipid peroxidation(Reference Azzi, Boscoboinik and Marilley183) and supplementation studies with vitamin E have indicated no significant effect on lipid peroxidation in vivo in healthy individuals(Reference Meagher, Barry and Lawson184, Reference Patrignani, Panara and Tacconelli185). These results dispute the role of vitamin E supplementation in individuals with normal baseline levels of antioxidants and oxidative stress (who then appear to be ‘non-responders’). As the majority of trial participants meet their RDA of these so-called antioxidant vitamins and with none of the large clinical trials assessing baseline levels of oxidative stress it is possible that the inclusion of ‘non-responders’ dilute the overall beneficial effect that is seen with responders, accounting for the disappointing overall findings. Clinical trials targeting individuals with an abnormal antioxidant status have shown more consistent benefits(Reference Salonen, Nyyssonen and Kaikkonen109, Reference Fang, Kinlay and Beltrame133, Reference Boaz, Smetana and Weinstein134, Reference Singh, Niaz and Rastogi161), indicating a role for subgroup targeting.
The Cambridge Heart Antioxidant Study (CHAOS) trial(Reference Stephens, Parsons and Schofield135) concluded that there was a significant reduction in cardiovascular events with α-tocopherol supplementation. Brown(Reference Brown186) concluded that a large number of these patients had a 3·5-fold increase in frequency for a polymorphism in the endothelial NO synthase gene that made them more prone to endothelial dysfunction and of greater need for vitamin E, hence further supporting subgroup targeting. However, it is still hard to accurately identify individuals exposed to increased oxidative stress due to the lack of efficient biomarkers for oxidative stress.
Patients with cardiovascular risk factors are exposed to greater amounts of oxidative stress(Reference Meagher and Rader187), which contributes to endothelial dysfunction(Reference Block, Dietrich and Norkus188, Reference Panza, Casino and Kilcoyne189). The enhanced level of oxidative stress is partly due to their reduced dietary intake of these so-called antioxidant vitamins(Reference Singh, Niaz and Bishnoi190) and this could possibly be responsible for the increased rate of atherosclerosis seen in these patients. The use of vitamins could retard the development of cardiovascular risk factors and reduce the risk of CVD.
Vitamin supplementation has been shown to improve endothelium-dependent dilatation in smokers(Reference Heitzer, Just and Munzel191) and in hypercholesterolaemic(Reference Neunteufl, Kostner and Katzenschlager192), hypertensive(Reference Plantinga, Ghiadoni and Magagna193) and diabetic patients(Reference Skyrme-Jones, O'Brien and Berry194).
Subgroup 1: smokers
Smoking is associated with an increased progression of atherosclerosis(Reference Poredos, Orehek and Tratnik195) and of heart disease(Reference Glantz and Parmley196), possibly mediated by exposure to increased levels of oxidative stress(Reference Harats, Ben-Naim and Dabach197–Reference Morrow, Frei and Longmire199). In smokers, plasma ascorbic acid, α-tocopherol and β-carotene levels are significantly depleted(Reference Mezzetti, Lapenna and Pierdomenico200–Reference Faure, Preziosi and Roussel202), partly as a consequence of increased utilisation(Reference Ayaori, Hisada and Suzukawa203, Reference Frei, Forte and Ames204), reduced regeneration of ascorbic acid(Reference Lykkesfeldt, Loft and Nielsen205) and their poorer diet(Reference Stryker, Kaplan and Stein206). Smokers have also been shown to have a down-regulated enzymic antioxidant defence system with reduced levels of catalase and glutathione peroxidase(Reference Hemalatha, Venkatesan and Bobby207) making them further prone to oxidative damage.
Supplementation with a combination of vitamins re-establishes a normal antioxidant status(Reference Lykkesfeldt, Christen and Wallock208) and reduces oxidative stress(Reference Fuller, Grundy and Norkus209–Reference Kim and Lee211) in smokers. The Antioxidant Supplementation in Atherosclerosis Prevention (ASAP) study(Reference Salonen, Nyyssonen and Salonen112) concluded a greater anti-atherogenic benefit in smokers than non-smokers. These vitamins have also on the other hand been shown to mediate a pro-oxidant effect with increased levels of oxidative stress(Reference Yang and Lowe162) and the likelihood of this is enhanced in smokers(Reference Princen, van Poppel and Vogelezang45, Reference Truscott212). However, the use of a combination of vitamins excluding β-carotene may prevent the increased likelihood of pro-oxidant effects and the negative findings encountered in the Beta-Carotene and Retinol Efficacy Trial (CARET)(Reference Omenn, Goodman and Thornquist114) and the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) study(115).
Subgroup 2: obese and overweight patients
In obese and overweight individuals fat-soluble vitamins can potentially become retained in visceral tissue, which can cause reduced serum levels of these vitamins. It has been shown that obese children have significantly reduced circulating levels of vitamin E and β-carotene(Reference Strauss213, Reference Decsi, Molnar and Koletzko214) and reduced LDL β-carotene and vitamin E levels(Reference Kuno, Hozumi and Morinobu215) compared with normal-weight children. As a result these individuals may be more prone to oxidative stress with an increased likelihood of endothelial dysfunction and LDL oxidation. This can partly account for the increased risk of CVD in obese or overweight individuals and supports a therapeutic role for supplementation with these vitamins in these individuals. Vitamin E supplementation in obese and overweight individuals has been shown to improve the metabolic profile (HbA1c, serum malondialdehyde levels and erythrocyte glutathione peroxidase activity were reduced)(Reference Ble-Castillo, Cleva-Villanueva and Diaz-Zagova216), increase antioxidant levels and reduce pro-oxidants(Reference Ble-Castillo, Cleva-Villanueva and Diaz-Zagova216, Reference Sutherland, Manning and Walker217). Randomised controlled trials are required to assess whether this reduction in oxidative stress reduces the development of CVD in obese or overweight individuals.
As obesity is a major risk factor for CVD, a large number of individuals in secondary prevention trials are likely to be overweight or obese. As the majority of trials do not directly assess the antioxidative actions of these vitamins in vivo and obese individuals are prone to the retention of vitamins in adipose tissue, it is plausible that these vitamins do not mediate their predicted effect in these individuals, possibly partly accounting for the lack of benefit seen in trials. This further emphasises the importance of identifying accurate biomarkers to assess vitamins' antioxidant effectiveness in vivo.
Subgroup 3: hypercholesterolaemic patients
Studies have shown that hypercholesterolaemic individuals have increased plasma lipid peroxide levels(Reference Davi, Alessandrini and Mezzetti218, Reference Yalcin, Sabuncu and Kilinc219) and possess LDL that is more susceptible to oxidation(Reference Cominacini, Pastorino and Garbin220, Reference Lavy, Brook and Dankner221). As α-tocopherol activity has been shown to be inversely related to cholesterol content in plaques(Reference Parker, Sabrah and Cap222), these individuals are also prone to have a diminished antioxidant status. This suggests that these individuals would benefit from vitamin supplementation.
Vitamin supplementation has been shown to alter lipid profile, mediating a reduction in total cholesterol, TAG and LDL levels(Reference Rezaian, Taheri and Mozaffari223), and positively correlating with HDL levels in individuals without diagnosed disease(Reference Simon and Hudes224, Reference Hallfrisch, Singh and Muller225). The use of vitamins in hypercholesterolaemia can be questioned as they have been shown to blunt the beneficial effect of simvastatin/niacin(Reference Brown, Zhao and Chait120). However, α-tocopherol has been excluded as a potential cause of this response(Reference Singh, Otvos and Dasgupta226). Nonetheless vitamin E supplementation in hypercholesterolaemic patients has resulted in a small but significant decrease in HDL-cholesterol levels and therefore caution still needs to be taken in regards to vitamin E supplementation(Reference Leonard, Joss and Mustacich227).
Statins have been shown to reduce vitamin E, β-carotene and ubiquinol-10 levels(Reference Jula, Marniemi and Huupponen228) and it has therefore been suggested that they may worsen the antioxidant status. This could possibly be due to statins reducing the circulating LDL fraction and therefore the delivery of these vitamins. This fact further emphasises a probably beneficial role of vitamin supplementation in hypercholesterolaemic patients.
In recent studies it has been shown that patients taking 10 mg atorvastatin per d gain an increase in plasma level of vitamin E (+42 %; P < 0·01)(Reference Cangemi, Loffredo and Carnevale229) and dual therapy with vitamins and statins has appeared to provide greater cardiovascular protection than statins on their own(Reference Blum, Milman and Shapira230). The lack of negative interaction between these agents further emphasises a beneficial role of supplementation with these vitamins in hypercholesterolaemic patients.
Subgroup 4: hypertensive patients
It has also been hypothesised that oxidative stress plays a role in the pathogenesis of hypertension and hypertension-induced damage through reducing NO levels and inducing endothelial dysfunction(Reference Landmesser and Harrison231). Hypertensive patients have been shown to be exposed to increased levels of lipid peroxidation and to have abnormal antioxidant status(Reference Russo, Olivieri and Girelli232). Observational trials have shown an inverse correlation between fruit and vegetable intake(Reference Appel, Moore and Obarzanek233), serum levels of putative antioxidant vitamins(Reference Chen, He, Hamm and Batuman234, Reference Salonen, Salonen and Ihanainen235) and the development of high blood pressure. These vitamins have been shown in in vitro studies to play a role in the aetiology of hypertension by restoring NO activity and endothelial function(Reference Plantinga, Ghiadoni and Magagna193, Reference On, Kim and Sohn236, Reference Tse, Maxwell and Thomason237). Vitamin E (400 IU/d) and vitamin C (1000 mg/d) supplementation resulted in beneficial effects on endothelium-dependent vasodilatation and arterial stiffness(Reference Plantinga, Ghiadoni and Magagna193) and in significantly lower systolic, diastolic and mean arterial blood pressure levels compared with the placebo group(Reference Rodrigo, Prat and Passalacqua238). Hypertensive patients have been more consistently shown to possess a reduction in ascorbic acid levels than those of any other antioxidant vitamins(Reference Fotherby, Williams and Forster239), possibly indicating an advantage of vitamin C supplementation over the other vitamins in these patients. Dietary intake(Reference Ness, Khaw and Bingham240) and plasma levels(Reference Bates, Walmsley and Prentice241) of ascorbic acid have been inversely related to blood pressure in some studies but not all(Reference Czernichow, Bertrais and Blacher242, Reference Miller, Appel and Levander243); in view of the lack of long-term benefit(Reference Kim, Sasaki and Sasazuki244) further research is required.
Subgroup 5: diabetic patients
Diabetic patients are exposed to increased levels of lipid peroxidation(Reference Gopaul, Anggard and Mallet245) as a result of LDL glycation(Reference Sobal, Menzel and Sinzinger246) and their increased levels of the small dense LDL subfraction(Reference Anderson, Gowri and Turner247), contributing to their high risk of macrovascular complications(Reference Uusitupa, Niskanen and Siitonen248). Vitamin E supplementation with doses that are greater than 800 IU/d in type 1 and 2 diabetic patients have been shown to reduce the oxidisability of LDL(Reference Fuller, Chandalia and Garg249, Reference Reaven, Herold and Barnett250) and improve endothelial function(Reference Skyrme-Jones, O'Brien and Berry194, Reference Ting, Timimi and Boles251). Supplementation with high-dose α-tocopherol has been associated with a reduced incidence of CHD(Reference Costacou, Zgibor and Evans252) and microvascular complications(Reference Bursell, Clermont and Aiello253) in diabetic users compared with non-users. Supplementation with 400 IU vitamin E per d also resulted in a significant reduction in cardiovascular events compared with a placebo group(Reference Milman, Blum and Shapira254). However, its long-term effects have not been confirmed by trials(Reference Stanner, Hughes and Kelly255, Reference Lonn, Yusuf and Hoogwerf256), possibly due to the use of lower vitamin dosages than those that have indicated short-term benefit in the small supplementation studies. At this point the available data are still too sparse to suggest the recommendation of vitamins to diabetic patients and more emphasis should be placed on targeting other diabetic-associated atherogenic factors.
Subgroup 6: patients with end-stage renal failure
IHD remains a leading cause of death in end-stage renal failure patients. Vitamin supplementation has been beneficial to patients with end-stage renal failure(Reference Boaz, Smetana and Weinstein134, Reference Jha, Flather and Lonn257) through reducing their increased levels of oxidative stress(Reference Handelman, Walter and Adhikarla258, Reference Dasgupta, Hussain and Ahmad259). The Secondary Prevention with Antioxidants of Cardiovascular disease in End-stage renal disease (SPACE) trial showed a 70 % reduction in myocardial infarct rates in haemodialysis patients with pre-existing CVD when supplemented with high-dose vitamin E(124, Reference Jha, Flather and Lonn257).
Subgroup 7: cardiac transplant or acute myocardial infarction patients
Atherosclerosis is a major complication of transplantation that limits the prolonged benefit of the transplant(Reference Mullins, Cary and Sharples260).
Vitamin supplementation has been beneficial to cardiac transplant(Reference Fang, Kinlay and Beltrame133) and acute myocardial infarction patients(Reference Singh, Niaz and Rastogi161, Reference Jaxa-Chamiec, Bednarz and Drozdowska261) in reducing their increased levels of oxidative stress(Reference Pechan, Danova and Olejarova262, Reference Singh, Niaz and Sharma263), making these vitamins a possible novel treatment for improving survival in these patients. The Indian Experiment of Infarct Survival(Reference Singh, Niaz and Rastogi161) and the Myocardial Infarction and Vitamins (MIVIT) pilot(Reference Jaxa-Chamiec, Bednarz and Drozdowska261) trial both confirmed a role for these vitamins in preventing post-myocardial-infarction complications and cardiac events. They have also been implicated in reducing the rejection of allogenic grafts(Reference Slakey, Roza and Pieper264), further emphasising a role in transplant patients.
The increased risk of congestive heart failure in vitamin E-supplemented post-myocardial-infarction patients(Reference Lonn, Bosch and Yusuf125, Reference Marchioli, Levantesi and Macchia265) indicates, however, the need of further trials. The authors of these trials hypothesised that these negative findings were due to pro-oxidant generation mediated by vitamin E.
Subgroup 8: elderly individuals
Elderly individuals are exposed to increased levels of oxidative stress(Reference Patrignani, Panara and Tacconelli185). Cohort studies(Reference Buijsse, Feskens and Kwape102, Reference Buijsse, Feskens and Schlettwein-Gsell103, Reference Losonczy, Harris and Havlik266) and trials(Reference Lee, Cook and Gaziano108) have both indicated benefit with supplementation in the elderly. In a subgroup analysis of the Women's Health study, only individuals above the age of 65 years gained a reduction, of 26 %, in cardiovascular events(Reference Lee, Cook and Gaziano108). The Atherosclerosis Risk in Communities (ARIC) study concluded an age-relationship between dietary intake and carotid atherosclerosis, with supplementation only showing benefit in women above the age of 55 years(Reference Kritchevsky, Shimakawa and Tell267). This therefore suggests that vitamin supplementation would be of benefit to elderly individuals.
Discussion
Antioxidant research has so far failed to confirm a role for vitamin E, vitamin C and β-carotene in the primary or secondary prevention of CVD. A total of nine primary and eleven secondary prevention trials, including approximately 150 000 and 60 000 participants respectively, have been disappointing. If there is a role for these vitamins in CVD, why is it that we have not identified it through trials? It has long been known that a high intake of fruit and vegetables is associated with a reduced incidence of CVD and it was initially hypothesised that vitamin E, vitamin C and β-carotene were the fundamental protective components that mediated this effect; as a consequence a range of studies was initiated to confirm their role.
Observational studies
It has long been believed that observational studies show that a high dietary intake of these vitamins is associated with a reduced risk of CVD and that there is a discrepancy between these studies and trials. However, even though this is true in regards to vitamin E, this is not the case when it comes to vitamin C and β-carotene. Observational studies have shown an inverse correlation between dietary intake of vitamin E (about 4–8 mg/d) and the incidence of CHD(Reference Knekt, Ritz and Pereira93–Reference Knekt, Reunanen and Jarvinen95). But the majority of studies have indicated no benefit with increased dietary intake of vitamin C(Reference Ito, Kurata and Suzuki92, Reference Kushi, Folsom and Prineas94, Reference Rimm, Stampfer and Ascherio96, Reference Osganian, Stampfer and Rimm98, Reference Buijsse, Feskens and Kwape102) and β-carotene(Reference Knekt, Ritz and Pereira93, Reference Riemersma, Wood and Macintyre268, Reference Iannuzzi, Celentano and Panico269) and those that have indicated a beneficial role have not adjusted for vitamin E intake(Reference Knekt, Reunanen and Jarvinen95) and hence its effects. The lack of benefit with a high dietary intake of β-carotene (about 890–5500 μg/d) and vitamin C (about 50–170 mg/d)(Reference Knekt, Ritz and Pereira93) is suggestive that β-carotene and vitamin C are not the relevant protective components in fruit and vegetables, therefore making one question whether they have a protective role in CVD.
Randomised controlled trials
Vitamin E
The positive evidence achieved with vitamin E in observational studies has led to more emphasis being put on vitamin E supplementation in randomised controlled trials. However, these positive findings achieved with vitamin E have not been possible to reproduce in randomised controlled trials. The underlying reason behind this discrepancy is still unclear. Even though observational studies have indicated a protective role for dietary intake of vitamin E, the beneficial role of vitamin E supplementation in CVD has only been supported by three studies(Reference Rimm, Stampfer and Ascherio96, Reference Stampfer, Hennekens and Manson97, Reference Losonczy, Harris and Havlik266). Trials looking at the effect of vitamin E supplementation using dosages between 330 and 800 IU/d have not supported a protective role for vitamin E in CVD. It has been argued that the dosages of vitamin E used in trials are too high, which causes the loss of beneficial effect. However, the supplementation of vitamin E with lower doses of ≤ 4·91 IU/d(Reference Kushi, Folsom and Prineas94) has not indicated any benefits even though this is equivalent to the level of vitamin E that is achieved with dietary intake and that has shown benefits in observational studies. Results from the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) trial showed an increased incidence of haemorrhagic stroke in the vitamin E-supplemented group compared with the placebo group(115); however, a prospective cohort study including participants from this trial concluded that those with higher circulating α-tocopherol within the normal range had a significantly lower total and CVD mortality(Reference Wright, Lawson and Weinstein106).
Vitamin C
With regard to vitamin C, studies have shown that supplementation with >500 mg vitamin C per d is associated with a lower risk of CHD(Reference Knekt, Ritz and Pereira93, Reference Osganian, Stampfer and Rimm98), suggestive that higher doses of vitamin C are required to mediate protective effects. However, the secondary prevention trial Women's Antioxidant Cardiovascular Study (WACS)(Reference Cook, Albert and Gaziano117) concluded no independent benefit of 500 mg vitamin C per d on cardiovascular endpoints, therefore disputing its independent therapeutic role in CVD.
β-Carotene
Randomised controlled trials that have assessed the effect of single supplementation with 20–50 mg β-carotene per d have not only been disappointing but have even been shown to increase the risk of CVD(Reference Omenn, Goodman and Thornquist114, 115). The levels of β-carotene used in trials is about 10 000 times greater than the levels used in dietary intake and it is possible that β-carotene mediates pro-oxidant effects at these levels, accounting for the negative results achieved.
Combinations
The lack of benefit seen in trials with the independent supplementation of vitamin E, regardless of the dosage used, or β-carotene or vitamin C supplementation is highly suggestive that supplementation with a single vitamin does not provide any reduced risk of CVD. As fruit and vegetables contain a range of vitamins in potential symbiosis this suggests that to gain benefit from vitamin supplementation it is necessary to try to mimic this environment through using a ‘cocktail’ of vitamins. The majority of large-scale trials have used a ‘cocktail’ of vitamins that has included β-carotene (Table 2)(Reference Hercberg, Galan and Preziosi107, Reference Christen and Gaziano111, Reference Blot, Li and Taylor116, 119, Reference Brown, Zhao and Chait120, Reference Singh, Niaz and Rastogi161) and the majority of these(Reference Hercberg, Galan and Preziosi107, Reference Blot, Li and Taylor116, 119) have not shown a beneficial effect and even indicated a negative effect in supplemented patients in comparison with non-supplemented patients(119, Reference Brown, Zhao and Chait120). It can be hypothesised to be due to the pro-oxidant effects of β-carotene overriding the antioxidative effects mediated by the other vitamins in the supplementation. Trials that have assessed the role of a combination of vitamins excluding β-carotene have indicated a potential beneficial role in atherosclerosis(Reference Salonen, Nyyssonen and Salonen112, Reference Fang, Kinlay and Beltrame133) and on atherosclerotic cardiovascular events(Reference Arad, Spadaro and Roth163) but have not reduced the risk of CVD in large-scale trials(Reference Cook, Albert and Gaziano117). The Women's Angiographic Vitamin and Estrogen (WAVE)(Reference Waters, Alderman and Hsia118) and WACS(Reference Cook, Albert and Gaziano117) trials showed disappointing findings with the combination of vitamin E and vitamin C in a secondary prevention study. In the WAVE trial there was actually a non-significant increase in coronary stenosis and all-cause mortality in post-menopausal women with 15–75 % coronary stenosis who were supplemented with the combination of vitamins E and C, compared with the placebo group. The reason behind the disappointing findings can be three-fold. First, the target group in the WACS and WAVE trials were women and it has been shown through other studies that they do not benefit significantly from vitamin supplementation with regard to CVD(Reference Hercberg, Galan and Preziosi107, Reference Salonen, Nyyssonen and Salonen112). Second, these trials were secondary prevention trials while the two other trials(Reference Salonen, Nyyssonen and Salonen112, Reference Arad, Spadaro and Roth163) that showed positive results were primary prevention trials. The oxidative modification hypothesis and the findings from prospective studies have suggested a beneficial role for supplementation with these vitamins in primary prevention; however, this has not been confirmed by clinical trials in regards to clinical endpoints. A role for these vitamins in secondary prevention has been disputed, with the evidence pointing towards an increase in total mortality in supplemented individuals with late-stage atherosclerosis(Reference Singh, Niaz and Rastogi161, Reference Yang and Lowe162, Reference Jha, Flather and Lonn257). The WACS and WAVE trials could have potentially included individuals who suffered from late-stage atherosclerosis, causing the negative effects mediated by this to blunt the predicted positive effect, giving an overall neutral effect. The neutral outcome or increase in clinical endpoints seen with the combination of vitamins C and E in secondary prevention trials suggests that supplementation is not an effective therapy in pre-existing CVD. However, the benefits seen in primary prevention trials suggest that the combined vitamin C and E supplementation may play a preventive role in those without pre-existing CVD. Third, the results from the WACS and WAVE trials could indicate that these vitamins are not the protective components in fruit and vegetables, further minimising the hope of a protective role for these vitamins in CVD.
Subgroup targeting
Let us consider the possibility that these vitamins have an optimal dose beyond which further intake does not mediate additional protection against oxidative stress (Fig. 1) and hence does not reduce LDL oxidation further. Salonen et al. (Reference Salonen, Nyyssonen and Salonen270) showed that ex vivo oxidisability and levels of lipid peroxide products were some of the strongest predictors of a 3-year increase in carotid wall thickness, which further supports a role for lipid peroxidation in atherosclerosis. As women are exposed to fewer cardiovascular risk factors(Reference Jousilahti, Vartiainen and Tuomilehto271) and have higher baseline serum concentrations of vitamins(Reference Somogyi, Herold and Kocsis272) they may be exposed to less oxidative stress compared with men. Therefore a lower intake of these vitamins may be required to achieve the optimal effect for the maximum protection against lipid peroxidation in women compared with men, explaining the benefit achieved with dietary intake in women but not in men(Reference Knekt, Ritz and Pereira93, Reference Kushi, Folsom and Prineas94). Trials using pharmacological doses of vitamin E (about 330–800 IU/d) have shown a trend towards a reduction in the incidence of CHD in men but not in women(Reference Hercberg, Galan and Preziosi107, Reference Salonen, Nyyssonen and Salonen112) and it is possible that at these dosages men achieve an optimal effect while in women supplementation moves them further along the plateau phase. Therefore supplementation would provide greatest benefit to those furthest away from their optimal level such as smokers, diabetics, and cardiac transplant and elderly patients. While the major large trials assessing the role of these vitamins have shown them to lack a beneficial role in CVD, the smaller trials assessing subgroup targeting have indicated a beneficial role in patients with end-stage renal disease(Reference Boaz, Smetana and Weinstein134), cardiac transplant(Reference Fang, Kinlay and Beltrame133) and acute myocardial infarction(Reference Jaxa-Chamiec, Bednarz and Drozdowska261). The combination of vitamins C and E in a secondary prevention trial has only shown benefit on clinical endpoints when targeting individuals (cardiac transplant patients) who are exposed to demonstrably increased levels of oxidative stress(Reference Fang, Kinlay and Beltrame133). Therefore through exploring subgroup targeting further in large-scale trials we could find a therapeutic role for these vitamins in CVD.
Fig. 1 Illustration of a hypothesis for the putative protective mechanism of antioxidants. The hypothesis suggests that antioxidants reach an optimal effect at a specific antioxidant concentration and that in women (––) the optimal antioxidant effect is reached with a lower antioxidant intake, i.e. dietary intake, than in men (- - -) in whom supplementation is needed to reach this optimal effect. It can be hypothesised that this is due to the pre-existing antioxidant levels being lower in men than in women and men being exposed to increased levels of oxidative stress.
To finally come to a conclusion on the role of vitamins in CVD one should probably conduct a primary prevention trial, using a combination of vitamins with 800 IU vitamin E per d and vitamin C >500 mg/d exclusive of β-carotene, and targeting subgroups that will potentially gain the most benefit such as diabetics, smokers, etc.
Flavonoids, fibre and folic acid
Individuals who consume large amounts of vitamins are less likely to smoke, have higher physical activity, are of higher socio-economic status(Reference Lyle, Mares-Perlman and Klein273) and more likely to consume other vitamins and to eat less saturated fat(Reference Reinert, Rohrmann and Becker274). A high intake of these vitamins could therefore act as a marker for other dietary or non-dietary factors, explaining the lack of benefits seen in trials. For example, a high dietary intake of fibre has been associated with a relative risk of 0·77 (95 % CI 0·61, 1·00) for CHD in observational studies(Reference Pereira, O'Reilly and Augustsson275–Reference Rimm, Ascherio and Giovannucci279). Flavonoids have shown an inconsistent role in CHD, with observational studies indicating a reduction in CHD mortality in those with a higher dietary intake of flavonoids compared with those with a lower dietary intake(Reference Geleijnse, Launer and Van der Kuip280, Reference Hertog, Feskens and Hollman281), while other studies have not indicated any beneficial role in CHD(Reference Lin, Rexrode and Hu282–Reference Rimm, Katan and Ascherio286). A high dietary intake of isoflavones has been shown to be associated with a reduced incidence of cerebral and myocardial infarction in women(Reference Kokubo, Iso and Ishihara287), possibly through its ability to reduce the progression of atherosclerosis(Reference Mursu, Nurmi and Tuomainen288). The role of flavonoids and fibre in CVD has not yet been assessed in randomised controlled trials. A high dietary intake of folic acid has been associated with a reduced incidence of CHD in one study(Reference Rimm, Willett and Hu289); however, a meta-analysis of randomised controlled trials showed it to have a neutral effect on CVD(Reference Bazzano, Reynolds and Holder290). There is a great need to assess the role of fibre and flavonoids in large-scale trials to be able more accurately to identify the protective components in fruit and vegetables.
Non-antioxidative properties
The causative role of oxidative stress in atherosclerosis has not been confirmed by in vivo studies and could therefore be an epiphenomenon(Reference Steinberg and Witztum172). The identification of the ‘oxidative stress response to inflammation’ hypothesis(Reference Stocker and Keaney291) makes the likelihood of a causative role for oxidative stress in atherosclerosis less plausible, therefore also making the disease-preventing roles for antioxidants less likely. However, as previously emphasised, vitamins such as vitamin E and vitamin C have been shown to mediate additional effects beyond their antioxidative properties including anti-inflammatory effects through altering gene expression and acting on signalling pathways that are activated by oxidised LDL. Therefore if oxidative stress does not play a role in atherosclerosis, it is still acknowledged that atherosclerosis is an inflammatory disease, and through their anti-inflammatory properties these vitamins can potentially still have a major role in atherosclerosis and CVD. Therefore the beneficial effect seen with these vitamins in observational studies could be due to these non-antioxidative properties. The lack of benefit seen with supplementation could be as a result of the lack of cofactors that are potentially present in fruit and vegetables that consequently can result in these properties not being fulfilled in trials. This further emphasises that supplementation should not only be a combination of vitamins E and C but also the relevant minerals and vitamins present in fruit and vegetables. These propositions have not yet been confirmed in vivo but through exploring these properties in a clinical context a future novel role in disease prevention for these vitamins can potentially be identified.
Conclusion
To resolve the discrepancy between observational studies and randomised clinical trials the design of the study has been the main alteration, either by increasing participant size, trial duration or type of supplementation, but this has left us empty handed. Through getting back to basic science and exploring whether oxidative stress has a causative role in atherosclerosis, a role for these vitamins in CVD will be further supported and also we will be enabled to define the optimal vitamin dose and type. The discovery of efficient and standardised oxidative biomarkers will enable the assessment of vitamins' antioxidant efficiency and the identification of individuals who would potentially be in greater need of vitamin supplementation. Future trials should look at the other components in fruit and vegetables, particularly flavonoids and fibre, to hopefully identify a novel preventative and therapeutic agent that can be used to prevent the rise in CVD around the world. The evidence is still insufficient to support a role for routine vitamin supplementation and at this stage more emphasis should be put in recommending a healthy lifestyle.
Acknowledgements
In the preparation of this paper there has been no conflict of interest. The article is funded by the Faculty of Medicine at the Imperial College School of Medicine, Science and Technology. S. H. was the main contributor to the article. M. S. amended the first and consequent drafts.
