Heart Protection Study Collaborative Group*

 

---

*Collaborators and participating hospitals are listed in Lancet 2002; 360: 7-22

 

---

Correspondence to: Heart Protection Study, Clinical Trial Service Unit and Epidemiological Studies Unit, Radcliffe Infirmary, Oxford OX2 6HE, UK (e-mail:hps@ctsu.ox.ac.uk)

 

Summary
Introduction
Patients and methods
Results
Discussion
References
 

 

Background It has been suggested that increased intake of various antioxidant vitamins reduces the incidence rates of vascular disease, cancer, and other adverse outcomes.

 

Methods 20 536 UK adults (aged 40-80) with coronary disease, other occlusive arterial disease, or diabetes were randomly allocated to receive antioxidant vitamin supplementation (600 mg vitamin E, 250 mg vitamin C, and 20 mg ß-carotene daily) or matching placebo. Intention-to-treat comparisons of outcome were conducted between all vitamin-allocated and all placebo-allocated participants. An average of 83% of participants in each treatment group remained compliant during the scheduled 5-year treatment period. Allocation to this vitamin regimen approximately doubled the plasma concentration of -tocopherol, increased that of vitamin C by one-third, and quadrupled that of ß-carotene. Primary outcomes were major coronary events (for overall analyses) and fatal or non-fatal vascular events (for subcategory analyses), with subsidiary assessments of cancer and of other major morbidity.

 

Findings There were no significant differences in all-cause mortality (1446 [14·1%] vitamin-allocated vs 1389 [13·5%] placebo-allocated), or in deaths due to vascular (878 [8·6%] vs 840 [8·2%]) or non-vascular (568 [5·5%] vs 549 [5·3%]) causes. Nor were there any significant differences in the numbers of participants having non-fatal myocardial infarction or coronary death (1063 [10·4%] vs 1047 [10·2%]), non-fatal or fatal stroke (511 [5·0%] vs 518 [5·0%]), or coronary or non-coronary revascularisation (1058 [10·3%] vs 1086 [10·6%]). For the first occurrence of any of these "major vascular events", there were no material differences either overall (2306 [22·5%] vs 2312 [22·5%]; event rate ratio 1·00 [95% CI 0·94-1·06]) or in any of the various subcategories considered. There were no significant effects on cancer incidence or on hospitalisation for any other non-vascular cause.

 

Interpretation Among the high-risk individuals that were studied, these antioxidant vitamins appeared to be safe. But, although this regimen increased blood vitamin concentrations substantially, it did not produce any significant reductions in the 5-year mortality from, or incidence of, any type of vascular disease, cancer, or other major outcome.

 

Lancet 2002; 360: 23-33

 

 

LDL-cholesterol may be rendered more atherogenic by oxidative modification that allows it to accumulate in the artery walls, and antioxidants have been shown to slow the progression of atherosclerosis in animal studies.^1-4 Vitamin E is a major antioxidant in LDL particles, and supplementation with vitamin E substantially prolongs the in-vitro resistance of LDL particles to oxidative damage, and has other potentially protective effects.^3-8 ß-carotene, which can also function as a fat-soluble antioxidant in certain physiological circumstances, is carried with vitamin E in the fatty core of LDL particles.^3,4 Vitamin C is a major water-soluble antioxidant in the plasma, and it can help to regenerate oxidised vitamin E.^4,9-11 In several non-randomised observational studies in different populations, dietary intake or plasma concentrations of these antioxidant vitamins were inversely associated with vascular disease incidence and mortality,^12-17 and blood concentrations of autoantibodies to oxidised LDL and the degree of LDL susceptibility to oxidative damage have been associated with atherosclerosis.^18,19 Dietary intake of antioxidant vitamins has also been reported in observational studies to be inversely associated with the incidence of various types of cancer.^16,17,20-23 But, without large-scale randomised evidence, the possibility that these associations merely reflect the effects of other aspects of the diet or lifestyle on disease rates cannot be ruled out.^20,24

 

Promising results on the progression of atherosclerosis^25,26 and on the incidence of vascular disease^27,28 have been reported from some small randomised trials of a few years of vitamin E in people with pre-existing vascular disease. But, the available results from much larger randomised trials of several years of vitamin E have been unpromising.^29-34 Similarly, the results thus far available from large long-term randomised trials of ß-carotene and of vitamin C have not provided good evidence of benefit.^29,30,34-37 Indeed, the results of some trials have even suggested that these vitamins have adverse effects (in particular on the incidence of haemorrhagic stroke and particular cancers^30,35), although this observation has not been confirmed by other trials.

 

The Heart Protection Study provides further evidence about the effects of these three antioxidant vitamins on vascular and non-vascular mortality and major morbidity by assessing 5 years of their supplementation in a large number of high-risk individuals.

 

 

Details of the study objectives, design, and methods are reported elsewhere^38,39 (including the protocol on the study website: www.hpsinfo.org) and are summarised below. As well as comparing antioxidant vitamins versus matching placebo in 20 536 randomised participants (which is the subject of the present report), a "2x2 factorial" design was used to allow the separate assessment of cholesterol-lowering therapy (see accompanying report³⁹).

 

Medical collaborators from 69 UK hospitals appointed senior nurses to run special clinics for the study (see Acknowledgments section in accompanying report³⁹), and obtained local ethics committee approval. Men and women aged about 40 years to 80 years with non-fasting blood total cholesterol concentrations of at least 3·5 mmol/L were eligible provided they were considered to be at substantial 5-year risk of death from coronary heart disease because of a past medical history of coronary heart disease, of other occlusive arterial disease, of diabetes mellitus, or of treated hypertension alone.^38,39 People were ineligible if they had other life-threatening conditions, such as chronic liver disease, severe renal disease, severe heart failure, severe chronic airways disease, or diagnosed cancer (other than non-melanoma skin cancer). In addition, anyone already taking high-dose vitamin E supplements, or in whom such supplements were considered indicated, was not to be randomised. Those individuals who appeared eligible for the study were given detailed information about it, and asked for their written agreement to participate. Potentially eligible people entered a prerandomisation "run-in" phase, which involved about 2 months of active vitamins. Compliant individuals who did not have a major vascular event or other serious problem during the run-in, and agreed to participate in the study for several years, were then randomly allocated to receive the antioxidant vitamins (600 mg synthetic vitamin E, 250 mg vitamin C, and 20 mg ß-carotene daily) or matching placebo capsules in specially prepared calendar packs. The central telephone randomisation system used a minimisation algorithm⁴⁰ to balance the treatment groups with respect to eligibility criteria and other major prognostic factors.

 

Following randomisation between July, 1994, and May, 1997, participants were to be seen in the study clinics for routine follow-up checks at 4, 8, and 12 months, and then 6-monthly until the final follow-up visits between May, and October, 2001. Those who became unable or unwilling to attend the clinics were to be contacted by telephone at the time of their scheduled follow-up (or, alternatively, follow-up was to be maintained via their general practitioner), and could continue to be supplied with their allocated study vitamins or matching placebo capsules by mail. Compliance with study treatment was assessed at each follow-up by reviewing the calendar-packed capsules remaining and, for those who had stopped, the reasons for doing so were sought. To assess the effects of the treatment allocation on blood concentrations of the vitamins being studied, assays were performed in non-fasting samples collected from about 5% of participants at the initial screening visit and at an average of about 3 years of follow-up (ie, the approximate mid-point of the study). Blood lipids were assessed in a selected sample of about 5% of participants due for follow-up at about the same time each year, and in all participants attending follow-up between August, 2000, and February, 2001. Differences in blood vitamin and lipid concentrations were based on comparisons between all those allocated the study vitamins and all those allocated placebo, irrespective of whether or not they were still compliant (with any missing data imputed from the screening values, assuming non-compliance).

 

Information was recorded at each follow-up of any suspected myocardial infarction, stroke, vascular procedure, cancer, or other serious adverse experience, and of the main reasons for all other hospital admissions (including day cases). Further details were sought from the participant's general practitioner (plus, if considered necessary, from any relevant hospital records) about all reports that might relate to major vascular events, cancers, or deaths, and from the UK national registries about the sites of any registered cancers and the certified causes of any deaths. All such information was reviewed by coordinating centre clinical staff who were kept unaware of the study treatment allocation, and events were coded according to prespecified criteria.³⁹

 

The data analysis plan was prespecified either in the original protocol³⁸ or in amendments (see study website) made before any analyses of the effects of treatment on clinical outcomes were available to the Steering Committee. All comparisons involved logrank analyses of the first occurrence of particular events during the scheduled treatment period after randomisation among all those allocated the vitamins versus all those allocated matching placebo capsules (ie, they were "intention-to-treat" analyses).⁴¹ The logrank analysis yielded the average event or death rate ratio (with the proportional reduction in this ratio expressed as a percentage), and the test of statistical significance (two-sided p value). The primary comparisons were of the effects of allocation to the vitamins on "major coronary events" (defined as non-fatal myocardial infarction or death from coronary disease) and on fatal coronary heart disease. Secondary comparisons were of the effects: (i) on major coronary events, and on "major vascular events" (defined as major coronary events, strokes of any type, and coronary or non-coronary revascularisations), during the first 2 years and during the later years of scheduled treatment; and (ii) on non-fatal or fatal strokes of any type. Other secondary comparisons included the effects on major coronary events, and on major vascular events, in various subcategories of participants determined at study entry. Tests for heterogeneity or, if more appropriate, trend were to be used to assess whether the proportional effects observed in specific subcategories differed clearly from the overall effects (after due allowance for multiple comparisons). In addition, several tertiary outcomes were prespecified (including site-specific cancer, cerebral haemorrhage, vascular procedures, and hospitalisation for various causes), again with due allowance in interpretation to be made for the exploratory and, perhaps, data-dependent nature of these, and the many other, analyses that might be performed.⁴¹

 

Based on previous studies in similar populations, it was estimated that there would be about 3000 major coronary events and 5000 major vascular events among 20 000 such high-risk patients followed for an average of 5 years.³⁸ If so, and if the antioxidant vitamins reduced these event rates by at least 10%, then the study had an excellent chance of demonstrating such effects at convincing levels of statistical significance. There were also expected to be more than 1000 deaths from causes other than coronary disease and more than 1000 new cancers during the scheduled follow-up. Such numbers would allow reasonably reliable assessment of the 5-year effects of the vitamin supplementation not just on all-cause mortality but also on the main non-coronary causes of death and on the main types of cancer. Interim analyses of mortality and of other major events were reviewed at least annually by the independent Data Monitoring Committee, but the investigators remained unaware of the results until completion of the scheduled treatment period.

 

The study was designed, conducted, analysed, and interpreted by the investigators entirely independently of all funding sources.

 

 

 

A total of 20 536 individuals (15 454 men and 5082 women) were randomised (figure 1), with 5806 aged at least 70 years at study entry.³⁹ Previous myocardial infarction was reported by 8510 (41% of those randomised), some other history of coronary disease by 4876 (24%), and no history of coronary disease by 7150 (35%). Among the 7150 participants without diagnosed coronary disease, 1820 had cerebrovascular disease, 2701 had peripheral arterial disease, and 3982 had diabetes mellitus (with some having more than one of these three conditions), whereas among the 13 386 with coronary disease, 1460, 4047, and 1981, respectively, had these conditions (again, with some "non-additivity" of these groups). Although treated hypertension was recorded in 8457 (41%) participants, only 237 (1%) were included on the basis of hypertension alone. The large size of the study (and the use of minimisation) produced good balance between the treatment groups (see subcategory figures below).

 

Figure 1: Trial profile   Numbers lost to follow-up relate to those without information to the end of the scheduled treatment period for mortality (as well as morbidity) and for morbidity alone.

 

The mean duration of follow-up was 5 years for all randomised patients: 5·3 years for those who survived to the scheduled end of study treatment and about half that for those who did not (yielding 50 837 person-years among all those allocated the study vitamins and 50 948 among all those allocated matching placebo). Compliance at each follow-up was defined as at least 80% of the scheduled vitamin or placebo capsules having been taken since the previous follow-up. Similar percentages of participants remained compliant in each treatment group, with the average during the study being 83% (table 1). Compared with placebo, allocation to the study vitamins approximately doubled the average plasma concentration of -tocopherol, increased that of vitamin C by about one-third, and quadrupled that of ß-carotene (table 2). There were also small, but highly significant, increases in the measured values of plasma total cholesterol, LDL cholesterol, and triglycerides among those allocated the study vitamins (table 2).

 

---

Follow-up (years)	Vitamin-allocated	Placebo-allocated
1	9003/10 113 (89%)	8971/10 082 (89%)
2	8369/9859 (85%)	8419/9876 (85%)
3	7918/9585 (83%)	7978/9642 (83%)
4	7427/9296 (80%)	7436/9333 (80%)
5	5640/7278 (77%)	5632/7317 (77%)
Study average (SE)	83% (0·1)	83% (0·1)
For missing follow-up, non-compliance is assumed.
Table 1: Compliance with study vitamins (80% taken) during follow-up

---

 

---

	Mean (SE) plasma concentrations*
	Vitamin-	Placebo-	Difference
	allocated	allocated	(vitamin-placebo)
Vitamins (µmol/L)
-tocopherol	49·5 (0·6)	27·0 (0·2)	22·5 (0·6)
Ascorbic acid	58·9 (1·0)	43·2 (1·0)	15·7 (1·4)
ß-carotene	1·22 (0·03)	0·32 (0·01)	0·89 (0·03)
Lipids and lipoproteins
Total cholesterol	4·89 (0·017)	4·74 (0·017)	0·15 (0·024)
LDL cholesterol	2·82 (0·014)	2·74 (0·014)	0·08 (0·019)
HDL cholesterol	1·10 (0·005)	1·13 (0·005)	-0·03 (0·007)
Triglycerides	2·13 (0·020)	1·92 (0·018)	0·21 (0·027)
Apolipoprotein A1	1·083 (0·004)	1·063 (0·005)	0·021 (0·007)
Apolipoprotein B	1·022 (0·007)	0·970 (0·007)	0·051 (0·010)
*Intention-to-treat comparisons, with missing data inputed from initial pretreatment screening values; mmol/L for total, LDL, HDL, and triglycerides, and g/L for apolipoproteins.
Table 2: Average plasma concentrations of vitamins and lipids during follow-up

---

 

During the scheduled treatment period, allocation to the study vitamins was associated with a non-significant excess in all-cause mortality (1446 [14·1%] vitamin vs 1389 [13·5%] placebo deaths; death rate ratio [RR] 1·04; 95% CI 0·97-1·12: figure 2). This excess involved slight, and non-significant, adverse trends in the mortality attributed to coronary heart disease (664 [6·5%] vs 630 [6·1%]; RR 1·06; 95% CI 0·95-1·18), other vascular causes (214 [2·1%] vs 210 [2·0%]; RR 1·02; 95% CI 0·84-1·24), and non-vascular causes (568 [5·5%] vs 549 [5·3%]; RR 1·04; 95% CI 0·92-1·17). No significant differences were observed between the treatment groups in any of the prespecified categories of non-vascular mortality.

 

Figure 2: Effects of vitamin allocation on cause-specific mortality   Rate ratios (RRs) are plotted (black squares with area proportional to the amount of statistical information in each subdivision) comparing outcome among participants allocated vitamins to that among those allocated placebo, along with their 95% CIs (horizontal lines; ending with arrowhead when CI extends beyond scale). For particular subtotals and totals, the result and its 95% CI are represented by a diamond, with the RR (95% CI) and its statistical significance given alongside. Squares or diamonds to the left of the solid vertical line indicate benefit, and to the right of the line indicate harm, with the vitamins (but this would be conventionally significant [p<0·05] only if the horizontal line or diamond did not overlap the solid vertical line).

 

The non-significant excess of vascular mortality was not supported by any excess of non-fatal vascular events, so it may well have been largely or wholly due to chance. Taking non-fatal and fatal events together, no significant differences were observed between the treatment groups in the numbers of participants who had non-fatal myocardial infarction or coronary death (1063 [10·4%] vitamin-allocated vs 1047 [10·2%] placebo-allocated; RR 1·02; 95% CI 0·93-1·11), or had any coronary or non-coronary revascularisation procedure (1058 [10·3%] vs 1086 [10·6%]; RR 0·98; 95% CI 0·90-1·06: figure 3). Nor were there significant differences in the numbers who had a non-fatal or fatal stroke (511 [5·0%] vs 518 [5·0%]; RR 0·99; 95% CI 0·87-1·12: figure 3), or strokes of any particular type or severity (figure 4). In particular, there was no significant effect of the study vitamins on the numbers having a haemorrhagic stroke (51 [0·5%] vs 53 [0·5%]).

 

Figure 3: Effects of vitamin allocation on first major coronary event, stroke, and revascularisation (defined prospectively as "major vascular events")   Symbols and conventions as in figure 2. Analyses are of the numbers of participants with a first event of each type during follow-up (with non-fatal and fatal events also considered separately), so there is some non-additivity between different types of event. MI=myocardial infarction.

 

Figure 4: Effects of vitamin allocation on first stroke   Symbols and conventions as in figure 2. For stroke type, analyses are of the numbers of participants having a first ischaemic or a first haemorrhagic stroke (with 11 having both stroke types), whereas those reporting only strokes that could not be classified are given in the final row. (Haemorrhagic stroke includes subarachnoid haemorrhage: 13 vitamin-allocated versus 7 placebo-allocated.) For stroke severity, black squares relate to the most severe stroke that could be classified (so these categories are mutually exclusive). Open squares are used to indicate rate ratios for participants who had only strokes of unknown type or severity.

 

A more stringent test of whether there was any net effect of the study vitamins on vascular disease is provided by analyses of all major coronary events, strokes, and revascularisations considered together. Allocation to the study vitamins did not significantly affect the numbers of participants who had any of these "major vascular events" (2306 [22·5%] vitamin vs 2312 [22·5%] placebo; RR 1·00; 95% CI 0·94-1·06: figure 3). The large numbers of events on which this comparison is based allows reliable assessment of the effects of the study vitamins in different circumstances. No significant difference was observed between the treatment groups in the numbers of participants who had a first major vascular event either during the first 2 years of scheduled treatment or, after more prolonged treatment, during subsequent years (figures 5 and 6). Nor were significant differences found in any of the various prior disease categories included, or in any other subcategory of participants examined (figures 7 and 8).

 

Figure 5: Effects of vitamin allocation on first major vascular event during follow-up   Symbols and conventions as in figure 2. Analyses are of numbers of participants having a first event during each year of follow-up and of those still at risk of a first event at the start of each year.

 

Figure 6: Life-table plot of effects of vitamin allocation on percentages having major vascular events   See figure 5 for numbers of participants having a first event during each year of follow-up.

 

Figure 7: Effects of vitamin allocation on first major vascular event in different prior disease categories   Symbols and conventions as in figure 2. There is no overlap between participants in "Any CHD" and "No CHD" baseline disease categories, but within each of these categories there is some overlap (and, hence, some non-additivity). 2 test on one degree of freedom is given for heterogeneity between rate ratios in participants with any prior coronary heart disease (CHD) versus those with no prior CHD.

 

Figure 8: Effects of vitamin allocation on first major vascular event in different categories of participant   Symbols and conventions as in figure 2. 2 tests on one degree of freedom are given for heterogeneity between rate ratios within dichotomous categories and for trend within other categories (with value >3·84 equivalent to p<0·05 before making allowance for multiple comparisons). Lipid categories relate to measured values at the initial screening visit prior to starting any statin treatment.39 *Slightly elevated creatinine defined as 110 µmol/L for women and 130 µmol/L for men, but <200 µmol/L for both.

 

New primary cancers (excluding non-melanoma skin cancer) were diagnosed in 800 (7·8%) of the participants allocated vitamins compared with 817 (8·0%) of those allocated placebo (RR 0·98; 95% CI 0·89-1·08: figure 9), and were associated with death in 359 (3·5%) versus 345 (3·4%) participants (RR 1·04; 95% CI 0·90-1·21: figure 2). These differences were not significant, and nor were there significant differences between the treatment groups in the incidence of cancers in any particular body system (figure 9), or in non-melanoma skin cancer (only one of which was fatal). When cancer sites were more finely divided to investigate hypotheses raised in previous studies, there were still no clearly significant differences between the treatment groups (for example, lung: 160 [1·6%] vitamin vs 141 [1·4%] placebo: p=0·3; stomach: 66 [0·6%] vs 50 [0·5%]: p=0·1; prostate: 138 [1·8%] vs 152 [2·0%] men: p=0·4).

 

Figure 9: Effects of vitamin allocation on site-specific cancer incidence   Symbols and conventions as in figure 2. Analyses are of the numbers of participants developing cancer at each site (excluding recurrences or new cancers at the same site), so there is some non-additivity between cancers at different sites. *Not including non-melanoma skin cancer, which is given separately.

 

Neuropsychiatric disorders--It had been suggested that antioxidant vitamins (in particular, vitamin E) might slow cognitive decline,⁴² so the modified Telephone Interview for Cognitive Status (TICS-m) questionnaire⁴³ was administered to participants during their final follow-up. A TICS-m score below 22 out of 39 was prespecified to be indicative of some cognitive impairment, and (as described in the accompanying paper³⁹) this well validated test appeared to have good discriminatory ability. However, no significant difference was observed between the treatment groups in the percentages of participants classified as cognitively impaired (23·4% vitamin-allocated vs 24·4% placebo-allocated) or in mean TICS-m score (24·11 vs 24·02; difference 0·09 [SE 0·07]). Similar numbers of participants in each treatment group were reported to have developed dementia during follow-up (31 [0·3%] vs 31 [0·3%]) or to have some other psychiatric disorder (62 [0·6%] vs 65 [0·6%]).

 

 

Respiratory disease--In non-randomised observational studies, higher intake of antioxidant vitamins has been associated with lower rates of asthma and chronic obstructive pulmonary disease,^44-46 so respiratory function was assessed by spirometry in all those attending the final follow-up visit. No significant differences were observed between the treatment groups in forced expiratory volume during one second (FEV₁: 2·06 L vitamin-allocated vs 2·06 L placebo-allocated; difference 0·00 L [SE 0·01]) or in forced vital capacity (FVC: 2·83 L vs 2·82 L; difference 0·01 L [SE 0·01]). Nor were significant differences observed in the numbers of participants hospitalised for chronic obstructive pulmonary disease or asthma (149 [1·5%] vs 133 [1·3%]) or for any other non-neoplastic respiratory cause (641 [6·2%] vs 642 [6·3%]).

 

 

Fractures--Higher dietary intake and blood concentrations of vitamin C have been associated in observational studies with greater bone density and lower prevalence of fractures.⁴⁷ Tertiary comparisons were, therefore, prespecified of the effects of the allocated vitamins on fractures (excluding the few related to road-traffic accidents). No significant differences were observed in the numbers of participants having any such fractures (234 [2·3%] vitamin vs 237 [2·3%] placebo) or those that are particularly related to osteoporosis (ie, hip, wrist, or spine: 101 [1·0%] vs 99 [1·0%]).

 

 

Other outcomes--There did not appear to be any significant difference between the treatment groups in the numbers hospitalised for any other particular reason (even before making allowance for the exploratory nature of such analyses). In particular, there was no significant support for the suggestion from observational studies that antioxidant vitamin supplementation might prevent cataracts,^48,49 with 379 (3·7%) vitamin-allocated versus 418 (4·1%) placebo-allocated participants reporting cataract (p=0·2). That result is consistent with the lack of clear effects on cataract in the large randomised Age-Related Eye Disease (ARED) trial of a similar antioxidant regimen,³⁴ and in other randomised trials of antioxidant vitamins that have assessed this outcome.^50-54 No significant differences were observed between the treatment groups in blood pressure or bodyweight recorded at the final follow-up visit.

 

 

 

The results of the Heart Protection Study indicate that 5 years of daily supplementation with 600 mg vitamin E, 250 mg vitamin C, and 20 mg ß-carotene is safe. But, although this regimen substantially increased the plasma concentrations of these vitamins, no significant benefits were observed among the high-risk individuals that were studied. These results effectively rule out any substantial reductions--or, indeed, increases--in heart attacks, strokes, cancers, or other major adverse events during 5 years of use of these vitamins. For example, proportional decreases or increases of as little as 10% in the rate of major vascular events are both excluded by even a 99% CI (0·93-1·08). Factors to consider in interpreting these findings include the population studied, the vitamin dosages (and combination) tested, and the duration of treatment and follow-up.

 

The non-randomised observational studies that found a lower incidence of cardiovascular events to be associated with higher intakes of different antioxidant vitamins were chiefly of people without known coronary or other vascular disease, and it has been suggested that these vitamins might be protective only before occlusive disease has developed.^55,56 But, more than 7000 of the participants in the Heart Protection Study had no evidence of coronary disease prior to randomisation, and among them there was no evidence of benefit. Likewise, about 4500 older American adults who had not had a recent vascular event were randomised in the ARED placebo-controlled trial of a similar combination regimen of 400 IU synthetic vitamin E, 500 mg vitamin C, and 15 mg ß-carotene daily. But, despite nearly 500 deaths during the scheduled treatment period in that study, it too did not find any difference in mortality between the treatment groups.³⁴ Oxidative stress is increased in some types of people, such as those with diabetes, and it has been suggested that antioxidant vitamins might be particularly effective in them.^55,57 There was, however, no evidence of benefit either among all 6000 people with diabetes in the present study or among the 3000 with diabetes who had no occlusive arterial disease at entry--or, indeed, in any other category of participant considered.

 

The same is true for trials of the primary prevention of vascular disease with particular components of this antioxidant regimen. Among about 4500 randomised people without diagnosed vascular disease in the Primary Prevention Project (PPP), 300 mg of synthetic vitamin E daily did not appear to produce any improvement in cardiovascular outcomes.³³ Similarly, about 22 000 of the 29 000 randomised participants in the Alpha-Tocopherol Beta-Carotene (ATBC) trial were free of coronary disease at entry, but no significant improvements in cardiovascular mortality were observed either with 50 mg of synthetic vitamin E daily (albeit a relatively low dose) or with 20 mg of ß-carotene daily.³⁰ Several other large-scale randomised trials of between 15 mg and 30 mg daily of ß-carotene (either alone^36,37 or in combination with 30 mg vitamin E²⁹ or 25 000 IU vitamin A³⁵) have been done in people without known vascular disease, but none has provided any good evidence of benefit. One of those trials, which was done in 30 000 people from Linxian in China,²⁹ also involved the separate assessment of 120 mg vitamin C daily (combined with 30 µg molybdenum). It had been thought that supplementation with antioxidant vitamins might be particularly effective in Linxian because of subclinical deficiencies of such micronutrients in this population,⁵⁸ and because of the high rates of strokes and certain types of cancers.⁵⁸ But, although plasma concentrations of ß-carotene and of ascorbic acid were increased substantially in that trial, there were no clearly significant reductions in cancer incidence or in mortality from vascular or other causes. Hence, the available evidence from large-scale randomised trials of different antioxidant regimens in various populations does not suggest that the lack of any clear benefits has been due to the types of people studied.

 

The daily dose of 600 mg of vitamin E tested in the Heart Protection Study is greater than the amounts that have been associated, in non-randomised observational studies, with 20-40% proportional reductions in the incidence of coronary events.^4,12,14,15 The synthetic form of vitamin E (all-rac--tocopherol: 1·1 IU/mg) used in this trial is considered to be somewhat less bioavailable than the natural form (RRR [2R, 4'R, 8'R], or d--tocopherol: 1·5 IU/mg).^7,59,60 Despite this, the study regimen doubled plasma concentrations of -tocopherol, and the synthetic formulation has been shown to prolong substantially the in vitro resistance of LDL to oxidative modification^6,8 (with no evidence that concomitant simvastatin influences these effects⁶¹). Moreover, the randomised Heart Outcome Prevention Evaluation (HOPE) trial used 400 IU daily of natural vitamin E in over 9000 similar high-risk patients, and it too found no evidence of any beneficial effects on vascular or other outcomes.³¹ The daily doses of 20 mg ß-carotene and of 250 mg vitamin C tested in the present trial are also well above the levels of dietary consumption that have been associated with lower rates of vascular disease and cancer in observational studies.^{4,12-17,20-23} The ß-carotene dose produced a highly significant four-fold increase in plasma concentrations, and although vitamin C concentrations increased by only about one third, doses of vitamin C above 200 mg daily do not appear to produce much further increase in either blood concentrations or inhibition of lipid peroxidation.⁶² It has been suggested that such antioxidant vitamins may be more effective when, as typically occurs in the diet, they are taken together.^2,11,56 But, despite the combination of dosages that were within the ranges considered likely to be effective,^2,23,56 and good compliance with the allocated treatment producing substantial increases in blood concentrations, there was no evidence of benefit in the Heart Protection Study.

 

It has also been suggested that antioxidants may chiefly prevent the development of new atherosclerotic plaques (although, presumably, slowing the uptake of modified LDL cholesterol should also slow the growth of existing plaques) and, as a consequence, that any benefits might take many years to emerge.^55,56 Most large randomised trials of these vitamins have involved an average of only about 4-6 years of treatment, but no benefits emerged among the 22 000 randomised participants in the Physicians' Health Study even after 12 years of ß-carotene supplementation.³⁶ Several randomised trials, including the present one, have shown that a substantial reduction in vascular events emerges within just 1-2 years of lowering LDL cholesterol.^39,63,64 Hence, if long-term treatment with antioxidant vitamins can eventually produce substantial protection against vascular disease by rendering LDL particles less atherogenic, then significant benefits would have been expected during the 5-year duration of the Heart Protection Study. But, there was no suggestion that beneficial effects were beginning to emerge even during the later years of treatment with these vitamins (figure 5). Participants will, however, continue to be followed for several more years to determine whether, after an average of 5 years of vitamin supplementation, any delayed effects on subsequent vascular events, cancers, or other major outcomes do eventually emerge.

 

Based on the presumption that the likelihood of benefit outweighs any low probability of harm, daily supplementation with a few hundred mg of vitamin E (and with other vitamins) has been recommended for middle-aged and older people.^7,56 But, despite assessing the combined effects of several years of substantial daily doses of different antioxidant vitamins (including 600 mg of vitamin E) in a large number of high-risk people, the Heart Protection Study has not been able to demonstrate any benefits from such supplementation. Moreover, this antioxidant regimen produced small, but definite, increases in the measured values of plasma triglycerides and LDL cholesterol (table 2), with a 3% higher mean LDL cholesterol concentration corroborated by a 5% higher mean apolipoprotein B concentration. (Although very high concentrations of vitamin C can interfere with lipid measurements, they have been found to produce artefactually low, rather than high, values for cholesterol and triglycerides.⁶⁵) Blood lipid concentrations during treatment were not measured routinely in most previous large-scale trials of antioxidant vitamins, but trends similar to those in table 2 have been observed for triglycerides,^32,66 though not for total or LDL cholesterol,^30,32,34,66 in those trials that did.

 

Other large-scale randomised trials of different antioxidant regimens in different populations are in progress,^4,12,23 and long-term follow-up continues in several completed trials (including the present one). But, in light of the unpromising results during at least 5 years of treatment in several large randomised trials, the lower risks of vascular disease and cancer found in observational studies among people with higher intake of these antioxidant vitamins must have been largely or wholly artefactual (ie, due to other differences in lifestyle that were actually responsible for the lower risks). Hence, continued recommendation of supplementation with such vitamins is difficult to justify. Instead, the main emphasis should be on those treatments (eg, aspirin, statins, angiotensin-converting-enzyme inhibitors, ß-blockers, and antihypertensive therapy) and those behavioural changes (eg, increasing physical activity and, particularly, stopping smoking) that are definitely known to prevent heart attacks, strokes, and other adverse outcomes.

 

The collaborators and committees involved in the conduct of the MRC/BHF Heart Protection Study are listed in the accompanying article³⁹ (which also provides a conflict of interest statement). The study was funded by the UK Medical Research Council, the British Heart Foundation, Merck & Co (manufacturers of simvastatin) and Roche Vitamins (manufacturers of the vitamins).

 

 

 

1 Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. Beyond cholesterol: modification of low-density lipoprotein that increases its atherogenicity.  N Engl J Med  1989; 320: 915-24. [PubMed]

 

 

2 Steinberg D and Workshop Participants. Antioxidants in the prevention of human atherosclerosis. Summary of the proceedings of a National Heart, Lung, and Blood Institute Workshop: September 5-6, 1991, Bethesda, Maryland.  Circulation 1992; 85: 2338-44. [PubMed]

 

 

3 Diaz MN, Frei B, Vita JA, Keaney JF. Antioxidants and atherosclerotic heart disease.  N Engl J Med 1997; 337: 408-16. [PubMed]

 

 

4 Food and Nutrition Board, Institute of Medicine. Dietary reference intakes for vitamin C, vitamin E, selenium, and carotenoids: a report of the Panel on Dietary Antioxidants and Related Compounds. Washington: National Academy Press, 2000.

 

 

5 Dieber-Rotheneder M, Puhl H, Waeg G, Striegl G, Esterbauer H. Effect of oral supplementation with D--tocopherol on the vitamin E content of human low density lipoproteins and resistance to oxidation.  J Lipid Res 1991; 32: 1325-32. [PubMed]

 

 

6 Jialal I, Fuller CJ, Huet BA. The effect of -tocopherol supplementation on LDL oxidation: a dose-response study.  Arterioscler Thromb Vasc Biol 1995; 15: 190-98. [PubMed]

 

 

7 Weber P, Bendich A, Machlin LJ. Vitamin E and human health: rationale for determining recommended intake levels.  Nutrition 1997; 13: 450-60. [PubMed]

 

 

8 Princen HMG, van Duyvenvoorde W, Buytenhek R, et al. Supplementation with low doses of vitamin E protects LDL from lipid peroxidation in men and women.  Arterioscler Thromb Vasc Biol 1995; 15: 325-33. [PubMed]

 

 

9 Frei B, Stocker R, England L, Ames BN. Ascorbate: the most effective antioxidant in human plasma.  Adv Exp Med Biol 1990; 264: 155-63. [PubMed]

 

 

10 Jialal I, Grundy SM. Preservation of the endogenous antioxidants in low density lipoprotein by ascorbate but not probucol during oxidative modification.  J Clin Invest 1991; 87: 597-601. [PubMed]

 

 

11 Upston JM, Terentis AC, Stocker R. Tocopherol-mediated peroxidation of lipoproteins: implications for vitamin E as a potential antiatherogenic supplement.  FASEB J  1999; 13: 977-94. [PubMed]

 

 

12 Jha P, Flather M, Lonn E, Farkouh M, Yusuf S. The antioxidant vitamins and cardiovascular disease: a critical review of epidemiologic and clinical trial data.  Ann Intern Med  1995; 123: 860-72. [PubMed]

 

 

13 Gaziano JM, Manson JE, Branch LG, Colditz GA, Willett WC, Buring JE. A prospective study of consumption of carotenoids in fruits and vegetables and decreased cardiovascular mortality in the elderly.  Ann Epidemiol 1995; 5: 255-60. [PubMed]

 

 

14 Kushi LH, Folsom AR, Prineas RJ, Mink PJ, Wu Y, Bostick RM. Dietary antioxidant vitamins and death from coronary heart disease in postmenopausal women.  N Engl J Med  1996; 334: 1156-62. [PubMed]

 

 

15 Klipstein-Grobusch K, Geleijnse JM, den Breeijen JH, et al. Dietary antioxidants and risk of myocardial infarction in the elderly: the Rotterdam Study.  Am J Clin Nutr  1999; 69: 261-66. [PubMed]

 

 

16 Simon JA, Hudes ES, Tice JA. Relation of serum ascorbic acid to mortality among US adults.  J Am Coll Nutr 2001; 20: 255-63. [PubMed]

 

 

17 Khaw KT, Bingham S, Welch A, et al. Relation between plasma ascorbic acid and mortality in men and women in EPIC-Norfolk prospective study: a prospective population study.  Lancet 2001; 357: 657-63. [Text]

 

 

18 Salonen JT, Ylä-Herttuala S, Yamamoto R, et al. Autoantibody against oxidised LDL and progression of carotid atherosclerosis.  Lancet 1992; 339: 882-87. [PubMed]

 

 

19 Regnström J, Nilsson J, Tornvall P, Landau C, Hamsten A. Susceptibility to low-density lipoprotein oxidation and coronary atherosclerosis in man.  Lancet 1992; 339: 1183-86. [PubMed]

 

 

20 Peto R, Doll R, Buckley JD, Sporn MB. Can dietary beta-carotene materially reduce human cancer rates?  Nature 1981; 290: 201-08. [PubMed]

 

 

21 Block G, Patterson B, Subar A. Fruit, vegetables, and cancer prevention: a review of the epidemiological evidence.  Nutr Canc 1992; 18: 1-29. [PubMed]

 

 

22 Flagg EW, Coates RJ, Greenberg RS. Epidemiologic studies of antioxidants and cancer in humans.  J Am Coll Nutr  1995; 14: 419-27. [PubMed]

 

 

23 Greenwald P. Cancer chemoprevention.  BMJ 2002; 324: 714-18. [PubMed]

 

 

24 MacMahon S, Collins R. Reliable assessment of the effects of treatment on mortality and major morbidity, II: observational studies.  Lancet 2001; 357: 455-62. [Text]

 

 

25 Salonen JT, Nyyssönenk K, Salonen R, et al. Antioxidant supplementation in atherosclerosis prevention (ASAP) study: a randomized trial of the effect of vitamins E and C on 3-year progression of carotid atherosclerosis.  J Intern Med 2000; 248: 377-86. [PubMed]

 

 

26 Fang JC, Kinlay S, Beltrame J, et al. Effect of vitamins C and E on progression of transplant-associated arteriosclerosis: a randomised trial.  Lancet 2002; 359: 1108-13. [Text]

 

 

27 Stephens NG, Parsons A, Schofield PM, et al. Randomised controlled trial of vitamin E in patients with coronary disease: Cambridge Heart Antioxidant Study (CHAOS).  Lancet 1996; 347: 781-86. [PubMed]

 

 

28 Boaz M, Smetana S, Weinstein T, et al. Secondary prevention with antioxidants of cardiovascular disease in endstage renal disease (SPACE): randomised placebo-controlled trial.  Lancet 2000; 356: 1213-18. [Text]

 

 

29 Blot WJ, Li J-Y, Taylor PR, et al. Nutrition intervention trials in Linxian, China: supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population.  J Natl Cancer Inst 1993; 85: 1483-92. [PubMed]

 

 

30 The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers.  N Engl J Med 1994; 330: 1029-35. [PubMed]

 

 

31 The Heart Outcomes Prevention Evaluation Study Investigators. Vitamin E supplementation and cardiovascular events in high-risk patients.  N Engl J Med 2000; 342: 154-60. [PubMed]

 

 

32 GISSI-Prevenzione Investigators (Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto miocardico). Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: results of the GISSI-Prevenzione trial.  Lancet 1999; 354: 447-55. [Text]

 

 

33 Collaborative Group of the Primary Prevention Project (PPP). Low-dose aspirin and vitamin E in people at cardiovascular risk: a randomised trial in general practice.  Lancet 2001; 357: 89-95. [Text]

 

 

34 Age-Related Eye Disease Study Research Group. A randomized, placebo-controlled clinical trial of high-dose supplementation with vitamins C and E and beta-carotene for age-related cataract and vision loss: AREDS Report No 9.  Arch Ophthalmol  2001; 119: 1439-52. [PubMed]

 

 

35 Omenn GS, Goodman GE, Thornquist MD, et al. Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease.  N Engl J Med 1996; 334: 1150-55. [PubMed]

 

 

36 Hennekens CH, Buring JE, Manson JE, et al. Lack of effect of long-term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease.  N Engl J Med 1996; 334: 1145-49. [PubMed]

 

 

37 Greenberg ER, Baron JA, Karagas MR, et al. Mortality associated with low plasma concentration of beta carotene and the effect of oral supplementation.  JAMA 1996; 275: 699-703. [PubMed]

 

 

38 MRC/BHF Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol-lowering therapy and of antioxidant vitamin supplementation in a wide range of patients at increased risk of coronary heart disease death: early safety and efficacy experience.  Eur Heart J  1999; 20: 725-41. [PubMed]

 

 

39 Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20 536 high-risk individuals: randomised placebo-controlled trial.  Lancet 2002; 360: 7-22. [Text]

 

 

40 White SJ, Freedman LS. Allocation of patients to treatment groups in a controlled clinical study.  Br J Cancer 1978; 37: 849-57. [PubMed]

 

 

41 Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II Analysis and examples.  Br J Cancer 1977; 35: 1-39. [PubMed]

 

 

42 Foley DJ, White LR. Dietary intake of antioxidants and risk of Alzheimer disease: food for thought.  JAMA 2002; 287: 3261-63. [PubMed]

 

 

43 Prince MJ, Macdonald AM, Sham PC, Richards M, Quraishi S, Horn I. The development and initial validation of a telephone-administered cognitive test battery (TACT).  Int J Methods Psych Res 1999; 8: 49-57. [PubMed]

 

 

44 Troisi RJ, Willett WC, Weiss ST, Trichopoulos D, Rosner B, Speizer FE. A prospective study of diet and adult-onset asthma.  Am J Respir Crit Care Med 1995; 151: 1401-08. [PubMed]

 

 

45 Grievink L, Smit HA, Ocké MC, van 't Veer P, Kromhout D. Dietary intake of antioxidant (pro)-vitamins, respiratory symptoms and pulmonary function: the MORGEN study.  Thorax 1998; 53: 166-71. [PubMed]

 

 

46 Hu G, Cassano PA. Antioxidant nutrients and pulmonary function: the Third National Health and Nutrition Examination Survey (NHANES III).  Am J Epidemiol  2000; 151: 975-81. [PubMed]

 

 

47 Simon JA, Hudes ES. Relation of ascorbic acid to bone mineral density and self-reported fractures among US adults.  Am J Epidemiol  2001; 154: 427-33. [PubMed]

 

 

48 Knekt P, Heliövaara M, Rissanen A, Aromaa A, Aaran R-K. Serum antioxidant vitamins and risk of cataract.  BMJ 1992; 305: 1392-94. [PubMed]

 

 

49 Simon JA, Hudes ES. Serum ascorbic acid and other correlates of self-reported cataract among older Americans.  J Clin Epidemiol  1999; 52: 1207-11. [PubMed]

 

 

50 Sperduto RD, Hu T-S, Milton RC, et al. The Linxian cataract studies: two nutrition intervention trials.  Arch Ophthalmol 1993; 111: 1246-53. [PubMed]

 

 

51 Teikari JM, Virtamo J, Rautalahti M, Palmgren J, Liesto K, Heinonen OP. Long-term supplementation with alpha-tocopherol and beta-carotene and age-related cataract.  Acta Ophthalmol Scand 1997; 75: 634-40. [PubMed]

 

 

52 Chylack LT, Brown NP, Brow A, et al, for the REACT Group. The Roche European American Cataract Trial (REACT): a randomized clinical trial to investigate the efficacy of an oral antioxidant micronutrient mixture to slow progression of age-related cataract.  Ophthalmic Epidemiol  2002; 9: 49-80. [PubMed]

 

 

53 Robman LD, McCarty CA, Tikellis G, et al. VECAT study: the effect of vitamin E on the progression of lens opacities (preliminary results). IOVS 2001; 42: S508 (abstr).

 

 

54 Christen WG. Beta-carotene and age-related cataract in a randomized trial of US physicians. IOVS 2001; 42: S518 (abstr).

 

 

55 Steinberg D, Witztum JL. Is the oxidative modification hypothesis relevant to human atherosclerosis? Do the antioxidant trials conducted to date refute the hypothesis?  Circulation 2002; 105: 2107-11. [PubMed]

 

 

56 Willett WC, Stampfer MJ. What vitamins should I be taking, doctor?  N Engl J Med  2001; 345: 1819-24. [PubMed]

 

 

57 Sundaram RK, Bhaskar A, Vijayalingam S, Viswanathan M, Mohan R, Shanmugasundaram KR. Antioxidant status and lipid peroxidation in type II diabetes mellitus with and without complications.  Clin Sci 1996; 90: 255-60. [PubMed]

 

 

58 Chen J, Campbell TC, Li J, Peto R. Diet, life-style and mortality in China. Oxford: Oxford University Press, 1990 (updated on www.ctsu.ox.ac.uk).

 

 

59 Brigelius-Flohé, Traber MG. Vitamin E: function and metabolism.  FASEB J 1999; 13: 1145-55. [PubMed]

 

 

60 Hoppe PP, Krennrich G. Bioavailability and potency of natural-source and all-racemic -tocopherol in the human: a dispute.  Eur J Nutr 2000; 39: 183-93. [PubMed]

 

 

61 Keech A, Collins R, Peto R, et al. Vitamin E supplementation inhibits LDL oxidation in patients at risk of coronary disease taking simvastatin or placebo. XI Int Symp Drugs Lip Met 1992; 59 (abstr).

 

 

62 Levine M, Wang Y, Padayatty SJ, Morrow J. A new recommended dietary allowance of vitamin C for healthy young women.  Proc Natl Acad Sci  2001; 98: 9842-46. [PubMed]

 

 

63 Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian simvastatin survival study (4S).  Lancet 1994; 344: 1383-89. [PubMed]

 

 

64 Sacks FM, Tonkin AM, Shepherd J, et al, for the Prospective Pravastatin Pooling Project Investigators Group. Effect of pravastatin on coronary disease events in subgroups defined by coronary risk factors: the Prospective Pravastatin Pooling Project.  Circulation 2000; 102: 1893-900. [PubMed]

 

 

65 Benzie IFF, Strain JJ. The effect of ascorbic acid on the measurement of total cholesterol and triglycerides: possible artefactual lowering in individuals with high plasma concentration of ascorbic acid.  Clinica Chimica Acta 1995; 239: 185-90. [PubMed]

 

 

66 Omenn GS, Goodman GE, Thornquist M, Brunzell JD. Long-term vitamin A does not produce clinically significant hypertriglyceridemia: results from CARET, the ß-carotene and retinol efficacy trial.  Cancer Epidemiol Biomarkers Prev 1994; 3: 711-13. [PubMed]