Is There Any Hope for Vitamin E?

During the past 15 years, epidemiological,^{1 - 2} basic biological,^{3 - 5} and experimental studies on atherosclerosis have supported the hypothesis that antioxidants protect against atherosclerosis^{6 - 8} by limiting low-density lipoprotein oxidation in the arterial wall. This mechanism inhibits the pathological accumulation of cholesteryl ester in plaque via the macrophage scavenger receptor, a process that can cause plaque rupture and cardiovascular events.^{9 - 10} Similarly, biological mechanisms have been identified in carcinogenesis that may be blocked by antioxidants.^{11 - 14} In the past decade, a number of prospective, randomized, placebo-controlled, 3- to 6-year clinical trials have been published, testing the effect of vitamin E and other antioxidant vitamins or their combinations on clinical manifestations of cardiovascular disease and cancer.^{15 - 21} These trials have surprisingly yet consistently shown that commonly used antioxidant vitamin regimens (vitamins E, C, beta carotene, or a combination) do not significantly reduce overall cardiovascular events or cancer.

Why not? Were these studies not long enough? In this issue of JAMA, the Heart Outcomes Prevention Evaluation (HOPE) investigators report an extension of the 9541-patient HOPE Vitamin E trial roughly 2.5 years beyond its previously reported¹⁵ 4.5-year mean follow-up. In the 2.5-year extension of HOPE (HOPE-TOO),²² 174 of the original 267 centers continued an extended follow-up. From these centers, 3994 of the 7030 original study enrollees who were still alive elected to continue the randomized vitamin E/placebo drug assignment. After a mean of 7.2 years of follow-up, vitamin E did not significantly reduce the relative risk (RR) of total cancer incidence (RR = 0.96; 95% confidence interval [CI], 0.84-1.09; P = .50), of cancer death (RR = 0.96; 95% CI, 0.75-1.22; P = .75), of a composite of cardiovascular events including cardiovascular death, nonfatal myocardial infarction, and stroke (RR = 1.05; 95% CI, 0.95-1.16; P = .31), or of individual components of this composite end point.²² These findings of lack of benefit from vitamin E (natural source, 400 IU α-tocopheryl acetate) during the extended study are consistent with the original HOPE report¹⁵ and with recent meta-analyses.^{23 - 24} The compelling message of this important study within its limits of confidence and over a 7-year exposure is that moderately high-dose vitamin E does not reduce overall risk of cardiovascular disease or cancer for 50- to 75-year-old men and women with established cardiovascular disease or diabetes, to whom the findings of this trial directly apply.²⁵

A subgroup finding of significantly reduced lung cancer risk with vitamin E reached statistical significance (RR = 0.72; 95% CI, 0.53-0.98; P = .04) at 4.5 years,¹⁵ but was not significant over the full 7-year follow-up (RR = 0.78; 95% CI, 0.55-1.10; P = .16). Of greater concern, another subgroup finding in HOPE-TOO was a vitamin E–associated increased risk of heart failure incidence that appeared in a secondary end point analysis in the 4.5-year report (RR = 1.13; 95% CI, 1.01-1.26; P = .03) and persisted in the 7-year extended follow-up (RR = 1.19; 95% CI, 1.05-1.35; P = .007), as did the risk of hospitalization for heart failure (RR = 1.40; 95% CI, 1.13-1.73; P = .002). The investigators dismissed the 4.5-year lung cancer benefit as a likely type I (false-positive) statistical error in a subgroup analysis.

However, the heart failure finding was thought to be more credible because it was observed in all predefined manifestations of heart failure, was present at 4.5 years for all heart failure events (RR = 1.17; P = .02), in the analysis of all enrolled patients at 7.2 years (RR = 1.13; P = .03), and in the sensitivity analysis using only patients enrolled in the 174 clinics continuing randomized therapy and follow-up (RR = 1.19; P = .007). Vitamin E also increased the risk of the more serious event of hospitalization for heart failure in the 7.2-year sensitivity analysis (RR = 1.40; P = .002). These findings are somewhat less compelling than the P value would indicate, because they occur in a secondary end point subgroup and without adjustment for multiple comparisons. The HOPE investigators appropriately point out that prior studies that have not reported on heart failure should be examined for this end point to determine whether their data support the validity of this finding. Thus, this potentially important observation awaits further confirmation.

Why is this report important? First, by extending HOPE and adding to the growing list of neutral prospective vitamin E trials (HOPE,¹⁵ GISSI-IV,¹⁶ ATBC,¹⁷ HPS,²¹ HATS²⁰ ), this report effectively closes the door on the prospect of a major protective effect of long-term exposure to this supplement, taken in moderately high dosage, against complications of atherosclerosis and overall cancer incidence. Second, in doing so, HOPE-TOO reemphasizes the importance of controlled clinical trials for testing important hypotheses deriving from basic biological findings or from epidemiological observations. The latter can mislead; well-designed clinical trials rarely do. Third, HOPE-TOO allows physicians to educate their patients as with the following response to inquiries about vitamin E, “In nearly 68 000 patients studied to date, there is no compelling evidence that higher doses of vitamin E reduce cardiovascular risk or cancer; there are even some hints that vitamin E, in excess of normal daily intake, may slightly increase the risk of ischemic events or of heart failure. You may hear that vitamin E is a ‘natural,’ yet effective, way to prevent heart disease or cancer, but this has proven to be a false hope. You should not be misled into neglecting other proven methods of prevention.” Fourth, while current evidence does not support a general cancer-protective effect of vitamin E, a favorable effect for specific cancers has not been fully ruled out, in particular for lung, oropharyngeal,²² and prostate cancers,^{17 ,22} for which protective biological mechanisms have been described.^{11 ,13} SELECT,²⁶ an ongoing 7- to 12-year trial of vitamin E, selenium, neither, or both in 35 000 healthy men, is sufficiently powered to detect a 25% reduction in prostate cancer incidence.

The antioxidant vitamin enthusiasm of the 1990s, followed by the emergence of disappointing data from randomized controlled trials, reflects a healthy balance of basic and clinical science and an excellent model for the development and evaluation of potential preventive therapies for commonly fatal chronic diseases. Within 10 to 15 years, oxidative mechanisms for cardiovascular disease and cancer were discovered and elucidated, implications for antioxidant therapy were recognized, and the effectiveness of antioxidant vitamins was tested in 8 major trials with nearly 140 000 participants randomized. Despite the compelling finding of “no benefit” overall against cardiovascular disease or cancer, there are specific disorders that may benefit from supplemental antioxidant vitamins. One condition of high oxidative stress is advanced renal failure in patients undergoing hemodialysis.^{27 - 29} One small, short trial showed that vitamin E significantly reduced the risk of myocardial infarction and a composite cardiovascular end point.²⁹ Another disorder is age-related macular degeneration of moderate or greater severity, in patients older than 55 years.³⁰ In a placebo-controlled factorial design trial comparing an antioxidant regimen (400 IU vitamin E, 500 mg vitamin C, and 15 mg beta carotene), or zinc and copper, or their respective placebos,³⁰ treated patients experienced a reduced risk of progression to severe age-related macular degeneration and loss of visual acuity at 5 years, significant only when antioxidants and zinc/copper were combined. These supplements did not benefit groups with initially mild forms of age-related macular degeneration or prevent cataract development.

While the results of these antioxidant therapies appear to be disappointing overall, it is important to note that the trial results do not necessarily negate prevailing views that oxidation of low-density lipoprotein and of certain other cellular targets may contribute importantly to atherogenesis and carcinogenesis. Indeed, there is remarkably little evidence that vitamin E effectively inhibits low-density lipoprotein oxidation in healthy humans lacking signs of increased oxidative stress.³¹ These uncertainties raise doubts about the ability of vitamin E to augment antioxidant defense mechanisms in vivo and leave many questions about low-density lipoprotein oxidation and atherosclerosis unanswered. Indeed, physiologically relevant oxidative reactions, including high-density lipoprotein (HDL) oxidation, that are not blocked by vitamin E in vitro have been described recently; these mechanisms may lead to appropriately targeted antioxidant approaches.^{32 - 33}

HOPE-TOO²² suggests that vitamin E significantly increases heart failure and results in subtle but nonsignificant trends toward increased ischemic complications of atherosclerosis.^{20 - 22} Indeed, a recent meta-analysis finds a significant dose-dependent adverse mortality effect of vitamin E when taken at 400 IU daily and above (RR = 1.04; P = .04).²⁴ Is there a biologically plausible explanation for these observed adverse effects of vitamin E? One hypothesis (Figure) is that expression of genes affecting lipoprotein metabolism³⁴ could be modified due to interference with the 9-cis retinoic acid–RXR interaction³⁵ by other isomers of retinoic acid, or by the effect of antioxidants on the enzymatically determined cellular oxysterol concentration.³⁵ A combination of vitamins E, C, and beta carotene actually decreases cardioprotective HDL₂ cholesterol levels (median 22%), and substantially blunts the increase in these levels associated with niacin and simvastatin therapy.^{20 ,36} HDL₂ reduction could be a mechanism for increasing atherosclerosis. Among those given vitamins in the Heart Protection Study,²¹ the average HDL (HDL₂ was not measured) was 3% lower, and the average total cholesterol was comparably higher, relative to placebo, effects consistent with a 4% to 5% increase in atherosclerotic cardiovascular events.³⁷ An increased total cholesterol/HDL ratio could also provide an explanation for the observed vitamin E–heart failure effect. Adequate myocardial perfusion depends on normal endothelial-dependent vascular reactivity, which is unfavorably affected by increases in total cholesterol/HDL.³⁸

The hopes for vitamin E alone or in combination with vitamin C and beta carotene have been diminished by a compelling body of clinical trial evidence and by certain adverse effects with plausible biological explanation. These hopes are now confined to modest expectations for specific disorders and there are concerns about adverse effects. While there is solid evidence linking oxidative processes to human disease (as well as to normal biological function), the details of these processes and of proposed therapeutic or preventive interventions appear to need considerable rethinking.