Risk of Mortality with Vitamin E Supplements: The Cache County Study

Article Outline

• Abstract
• Methods
  • Statistical Analyses
• Results
• Discussion
• Acknowledgments
• References
• Copyright

Abstract 

Purpose

A recent meta-analysis reported increased mortality in clinical trial participants randomized to high-dose vitamin E. We sought to determine whether these mortality risks with vitamin E reflect adverse consequences of its use in the presence of cardiovascular disease.

Methods

In a defined population aged 65 years or older, baseline interviews captured self- or proxy-reported history of cardiovascular illness. A medicine cabinet inventory verified nutritional supplement and medication use. Three sources identified subsequent deaths. Cox proportional hazards methods examined the association between vitamin E use and mortality.

Results

After adjustment for age and sex, there was no association in this population between vitamin E use and mortality (adjusted hazard ratio [aHR] 0.93; 95% confidence interval [CI], 0.74-1.15). Predictably, deaths were more frequent with a history of diabetes, stroke, coronary artery bypass graft surgery, or myocardial infarction, and with the use of warfarin, nitrates, or diuretics. None of these conditions or treatments altered the null main effect with vitamin E, but mortality was increased in vitamin E users who had a history of stroke (aHR 3.64; CI, 1.73-7.68), coronary bypass graft surgery (aHR 4.40; CI, 2.83-6.83), or myocardial infarction (aHR 1.95; CI, 1.29-2.95) and, independently, in those taking nitrates (aHR 3.95; CI, 2.04-7.65), warfarin (aHR 3.71; CI, 2.22-6.21), or diuretics (aHR 1.83; CI, 1.35-2.49). Although not definitive, a consistent trend toward reduced mortality was seen in vitamin E users without these conditions or treatments.

Conclusions

In this population-based study, vitamin E use was unrelated to mortality, but this apparently null finding seems to represent a combination of increased mortality in those with severe cardiovascular disease and a possible protective effect in those without.

Keywords: Vitamin E, Mortality, Cardiovascular disease, Cohort study, Proportional hazards, Survival analysis

Antioxidant vitamins are reputed to prevent several diseases of aging¹ and have been reported to reduce cardiac and all-cause mortality.² In the last 2 years, however, controversy has developed over whether vitamin E is protective against cardiovascular disease or cancer,³ or whether it puts individuals at increased risk for an adverse event, including death.⁴ A recent meta-analysis of 19 clinical trials showed increased death rates in participants randomized to vitamin E supplements at doses of 400 international units (IU) or higher per day.⁴ Most of the trial participants had serious cardiovascular disease or were otherwise at increased risk for cardiovascular events. It is therefore unclear whether vitamin E might produce lesser mortality or even have a beneficial effect in those at lower cardiovascular risk. To address this question, we examined the mortality associated with vitamin E use in a population-based study to see whether typical patterns of use (∼400 IU/day) showed different apparent effects in those with and without existing cardiovascular disease or risk factors.

Back to Article Outline

Methods 

The Cache County Study is a prospective cohort study of persons aged 65 years or older residing in Cache County, Utah, as of January 1, 1995. Study methods have been described in detail elsewhere.⁵ Briefly, 5092 persons (90% of those eligible) completed initial interviews at their place of residence. Most of these (97%) also provided buccal DNA for genotyping at APOE, the polymorphic genetic locus for apolipoprotein E. Some 3 years later, participants were re-interviewed.

In the initial (baseline) 2-hour interview, participants were asked about lifetime history of illness and current use of medicines or nutritional supplements, as verified by visual inspection of medicine cabinet contents. A history of hypertension, hypercholesterolemia, diabetes mellitus, stroke, coronary artery bypass graft surgery, or myocardial infarction was recorded when participants reported a physician’s diagnosis or treatment for these conditions. Sustained use of vitamin E was noted if participants reported use of 4 or more doses weekly of any vitamin supplement containing 400 IU or more. We identified deaths through January 25, 2002, by using local newspaper obituaries and county-specific data from the Utah Department of Vital Statistics.⁶ Independent mortality data also were obtained from the Utah Population Database,⁷ which monitors vital statistics over the entire state. Capture–recapture analyses⁸ suggested that the combination of these methods detected 96.3% (95% confidence interval [CI], 95.71-96.8) of all deaths among those originally resident in the county. The institutional review boards of Utah State University, Duke University, and Johns Hopkins University approved the protocol and consent forms, and participants gave informed consent for all procedures.

Statistical Analyses 

Baseline demographic characteristics of vitamin E users and non-users were evaluated using chi-square statistics for categoric variables and t tests for continuous measures. In a series of models, Cox proportional hazards methods⁹ were used to evaluate the relationship between vitamin E use and mortality. Improvement in model fit with additional terms was assessed using the likelihood ratio chi-square statistic. Additional analyses were conducted using orthogonal variables for vitamin E use and selected cardiovascular diseases. Orthogonal models allow for the direct comparison of risks for non-vitamin E users without disease (−, −), with vitamin E users without disease (+,−), non-users with disease (−,+), and vitamin E users with disease (+,+). All analyses used the PHREG routine in SAS version 8.0 (SAS Institute, Cary, NC) and the SAS EXACT method to account for ties on a continuous time scale.

Back to Article Outline

Results 

Of the 5092 initial participants, 355 were excluded here because they had prevalent dementia.⁵ Another 321 were excluded because of missing data on supplement use (n=31), medical history (n=180), or APOE genotype (n=121). Among the remaining analysis sample of 4416 participants, 426 individuals (10%) reported sustained use of vitamin E at baseline. Such use was typically habitual (most reported having used the vitamin for at least 2 years), and 66% of these who survived reported use at the follow-up interview. Table 1 presents baseline demographic characteristics of the sample categorized by baseline vitamin E use. Vitamin E users were younger and better educated than non-users. Unexpectedly, there was significantly more vitamin E use in participants with at least one APOE ε4 allele.

Table 1. Baseline Demographic Characteristics of 4416 Residents of Cache County, Utah

Characteristic	Vitamin E Users	Nonusers	Total
N (%)	426(9.7)	3990(90.3)	4416
Mean baseline age (SD)⁎	73.8(6.1)	75.2(7.0)	75.1(6.9)
Female (%)	247(58.0)	2260(56.6)	2507(56.8)
Mean education level (SD)†	13.5(2.9)	13.2(2.9)	13.3(2.9)
Number of APOE ε4 alleles
0 APOE ε4 alleles (%)	274(64.3)	2788(69.9)	3062(69.3)
1 APOE ε4 allele (%)	141(33.1)	1107(27.7)	1248(28.3)
2 APOE ε4 alleles (%)	11(2.6)	95(2.4)	106(2.4)
Cardiovascular history
Stroke (%)	13(3.1)	202(5.1)	215(4.9)
Diabetes (%)†	37(8.7)	505(12.7)	542(12.3)
CABG (%)	36(8.5)	278(7.0)	314(7.1)
MI (%)	52(12.2)	516(12.9)	568(12.9)
Deaths (%)†	88(20.7)	1041(26.1)	1129(25.6)

APOE=apolipoprotein E; CABG=coronary artery bypass graft surgery; MI=myocardial infarction; SD=standard deviation.

Values are number (percent) unless indicated as mean (SD).

We identified 1129 deaths over the subsequent 7-year observation period. Table 2 shows a series of proportional hazards models relating mortality with the use of vitamin E and several other variables. The crude hazard ratio for mortality with vitamin E was 0.77 (CI, 0.62-0.96), but this association was evidently confounded with age, sex, and cardiovascular disease. Inclusion of age and sex in Model 1 yielded an adjusted hazard ratio (aHR) of 0.93 (CI, 0.74-1.15). Models 2 to 8 show expected increases in mortality in participants with a history of stroke, myocardial infarction, or coronary artery bypass graft surgery, and, independently, with use of several medicines commonly prescribed for cardiovascular disease. However, addition of these terms as covariates produced no notable change in the near-null association of vitamin E use and mortality. In contrast, a series of orthogonal models in Table 3 suggested significant mortality risks in vitamin E users who had a history of stroke, myocardial infarction, or coronary bypass graft surgery. Notable incremental risks were also evident with vitamin E use among those taking nitrates or warfarin, with a suggestion of increased risk among those taking diuretics. These models differ from those presented in Table 2 in that the comparison group is a reference category composed of individuals who do not use vitamin E and do not report a history of each of the listed cardiovascular conditions. A final proportional hazards model that included all these main effects, as well as diabetes, and their interactions with vitamin E yielded an aHR of 0.69 (CI, 0.49-0.97) for vitamin E users with none of these conditions or treatments (data not shown). Vitamin E use showed no notable interaction with angiotensin-converting enzyme inhibitors, beta-blockers, or calcium channel blockers, nor was there an interaction with aspirin as a predictor of mortality (data not shown). Separate analyses did not suggest that individuals with severe cardiovascular disease were more frequent users of vitamin E supplements than others (odds ratio [OR] with stroke 0.68 CI, 0.37-1.17; OR with coronary bypass graft surgery 1.32, CI, 0.85-1.99; OR with myocardial infarction 0.92, CI, 0.64-1.31).

Table 2. Multivariable Adjusted Cox Proportional Hazard Ratios for Mortality Among 4416 Elderly Residents of Cache County, Utah, in Models of Increasing Complexity

Characteristic	Crude HR (95% CI)	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Vitamin E	0.77(0.62-0.96)	0.93(0.74-1.15)	0.96(0.77-1.19)	1.0(0.80-1.24)	0.99(0.80-1.24)	1.00(0.80-1.25)	1.00(0.81-1.25)	1.02(0.82-1.27)	1.00(0.80-1.25)
Baseline age (y)	1.11(1.10-1.11)	1.11(1.10-1.12)	1.11(1.10-1.11)	1.11(1.10-1.11)	1.11(1.10-1.11)	1.11(1.10-1.12)	1.11(1.10-1.12)	1.11(1.10-1.12)	1.10(1.10-1.11)
Female sex	0.76(0.68-0.86)	0.67(0.59-0.75)	0.67(0.60-0.76)	0.69(0.61-0.77)	0.71(0.63-0.80)	0.73(0.65-0.82)	0.73(0.64-0.82)	0.73(0.64-0.82)	0.68(0.60-0.77)
Stroke	2.81(2.32-3.41)		2.23(1.83-2.71)	2.09(1.72-2.54)	2.03(1.67-2.47)	2.02(1.66-2.46)	1.97(1.61-2.40)	1.77(1.45-2.17)	1.68(1.37-2.06)
Diabetes mellitus	2.02(1.74-2.34)			1.91(1.64-2.21)	1.84(1.59-2.14)	1.84(1.58-2.13)	1.81(1.56-2.10)	1.81(1.56-2.10)	1.70(1.46-1.98)
MI	1.77(1.52-2.05)				1.36(1.17-1.59)	1.21(1.02-1.43)	1.16(0.97-1.38)	1.11(0.93-1.32)	1.08(0.90-1.28)
CABG	1.76(1.46-2.13)					1.49(1.20-1.85)	1.49(1.20-1.85)	1.41(1.14-1.75)	1.39(1.12-1.72)
Nitrates	2.81(2.19-3.61)						1.32(1.01-1.72)	1.25(0.96-1.64)	1.25(0.96-1.64)
Warfarin	2.39(2.02-2.83)							1.72(1.44-2.06)	1.59(1.33-1.90)
Diuretics	1.94(1.72-2.19)								1.45(1.28-1.65)
LR χ2, df		645.94,3	698.89,4	763.31,5	778.11,6	790.27,7	794.11,8	825.85,9	858.86,10
Difference in LR χ2, df, and P value⁎			52.95,df1,P<.0001	64.42,df1,P<.0001	14.8,df1,P<.0001	12.16,df1,P=.0004	3.84,df1,P<.05	31.74,df1,P<.0001	33.01,df1,P<.0001

CABG=coronary artery bypass graft surgery; CI=confidence interval; df=degrees of freedom; DM=diabetes mellitus; HR=hazard ratio; LR=likelihood ratio; MI=myocardial infarction.

Table 3. Orthogonal Adjusted Cox Proportional Hazard Models for Risk of Death Among 4416 Residents of Cache County, Utah

Multivariable Hazard Ratios⁎
Variables	n	aHR (95% CI)	Variables	n	aHR (95% CI)
No vitamin E, no stroke	3788	1.00(reference)	No vitamin E, no diuretics	2843	1.00(reference)
Vitamin E, no stroke	413	0.96(0.77-1.21)	Vitamin E, no diuretics	301	0.82(0.60-1.11)
No vitamin E, stroke	202	1.96(1.60-2.40)	No vitamin E, diuretics	1147	1.47(1.29-1.67)
Vitamin E, stroke	13	3.64(1.73-7.68)	Vitamin E, diuretics	125	1.83(1.35-2.49)
No vitamin E, no CABG	3712	1.00(reference)	No vitamin E, no nitrates	3882	1.00(reference)
Vitamin E, no CABG	390	0.81(0.63-1.04)	Vitamin E, no nitrates	413	0.93(0.74-1.17)
No vitamin E, CABG	278	1.28(1.01-1.61)	No vitamin E, nitrates	108	1.18(0.89-1.57)
Vitamin E, CABG	36	4.40(2.83-6.83)	Vitamin E, nitrates	13	3.95(2.04-7.65)
No vitamin E, no MI	3474	1.00(reference)	No vitamin E, no warfarin	3688	1.00(reference)
Vitamin E, no MI	374	0.86(0.70-1.12)	Vitamin E, no warfarin	401	0.91(0.72-1.16)
No vitamin E, MI	516	1.13(0.95-1.35)	No vitamin E, warfarin	302	1.63(1.36-1.97)
Vitamin E, MI	52	1.95(1.29-2.95)	Vitamin E, warfarin	25	3.71(2.22-6.21)
No vitamin E, no CVD†	2318	1.00(reference)
Vitamin E, no CVD†	254	0.71(0.49-1.03)
No vitamin E, any CVD†	1672	1.60(1.41-1.81)
Vitamin E, any CVD†	172	1.88(1.43-2.47)

aHR=adjusted hazard ratio; CI=confidence interval; CABG=coronary artery bypass graft surgery; MI=myocardial infarction; CVD=cardiovascular disease.

Finally, because many of the participants who used vitamin E supplements also used vitamin C (70%) and multivitamin supplements (51%), and because of the possibility of synergistic effects specifically between vitamin E and vitamin C,¹⁰ we reevaluated the association of vitamin E use and mortality after controlling for the use of these other supplements. This did not significantly change our findings (data not shown).

Back to Article Outline

Discussion 

The Cache County cohort of persons aged 65 years and older showed no overall change in mortality with vitamin E use, but this “null” relationship seemed to reflect opposite associations with mortality in those with and without a history of stroke, coronary bypass graft surgery, or myocardial infarction, or use of nitrates or warfarin. Although this dichotomy in the apparent effect of vitamin E may seem novel, it is in broad accord with a complex previous evidence base on vitamin E and mortality from published randomized trials.³, ¹¹, ¹², ¹³, ¹⁴, ¹⁵, ¹⁶ In general, these trial results suggest a null relationship of vitamin E and all-cause mortality¹¹, ¹², ¹³, ¹⁶; no evident protection against new-onset cardiovascular disease¹¹, ¹², ¹³, ¹⁴, ¹⁶; and an increase in cardiac death or overall mortality in those with preexisting cardiac disease.³, ¹⁴, ¹⁵ New findings here suggest a borderline significant 30% reduction in overall mortality in those without history of any preexisting cardiovascular disease or diabetes, and the vitamin’s evident harm to others may relate in part to specific and harmful interactions between the use of vitamin E and either nitrates or warfarin, both common treatments for those with severe cardiac illness.

The adverse interaction of vitamin E with warfarin may reflect the known effect of the vitamin on inhibition of vitamin K-dependent clotting.¹⁷, ¹⁸ The interaction with nitrates, especially evident here, is harder to understand, although some explanations are available.¹⁹, ²⁰, ²¹, ²²

Although our findings could potentially be explained by an artifact of increased use of vitamin E supplements (as a presumed preventive measure) among Cache County Study participants with cardiovascular disease, we doubt this explanation. Our participants with a history of stroke, coronary bypass graft surgery, or myocardial infarction were no more likely to report vitamin E use than those without these conditions.

Some other limitations in our methods must be acknowledged. First, these are observational data in a study that was not originally designed to test the hypotheses investigated here, so our findings require confirmation. Second, our results rely on self-report of both medication/supplement use and cardiovascular disease history, which may be subject to misclassification. Computerized pharmacy dispensing records would have been a stronger method of classification for drug exposures, although there is doubt whether this approach would provide rigorous data on vitamin supplement use. Likewise, medical diagnostic histories from a health maintenance organization or other systematically collected health records database would have provided more accurate health history data. Inaccurate methods of classification in observational studies increase the possibility of various forms of bias and confounding, including especially the effects of “noise”, biasing results toward the null. Although we find it difficult to conceive how misclassification error would have led spuriously to the present results, we cannot exclude that possibility. Third, we did not consider variability in the dosage of vitamin E in these analyses. Vitamin E use was dichotomized on the basis of regular use of 400 IU or more on a daily basis because this is the typical formulation available. Future analysis should consider more specific breakdown of dosages. Finally, the population under study is aged 65 years or older, primarily white, and confined to one geographic area in which most participants are members of the Church of Jesus Christ of Latter Day Saints (91%); thus, these results may not generalize to other populations.

We suggest that our results warrant further caution regarding use of vitamin E by those with existing cardiovascular disease or those taking nitrates or warfarin. If sustained, some of the associations noted here will have significant health implications, the more so given the increasing use of vitamin supplements in the United States and Europe. Further investigation of these relationships will therefore be important.

Back to Article Outline

Acknowledgments 

We thank Dr Bruce Psaty for his helpful comments on an earlier version of this article. Ken Smith, PhD, and the University of Utah Huntsman Cancer Institute provided mortality data from the Utah Population Database. We are grateful to the neurogenetics laboratory of the Bryan Alzheimer’s Disease Research Center at Duke University for the APOE genotyping.

Other Cache County Study of Memory, Health, and Aging investigators include James Anthony, PhD; Erin Bigler, PhD; Ron Brookmeyer, PhD; James Burke, MD, MPH; Eric Christopher, MD; Chris Corcoran, PhD; Jane Gagliardi, MD; Robert Green, MD; Michael Helms; Christine Hulette, MD; Liz Klein, MPH; Carol Leslie, MS; Lawrence Mayer, MD; John Morris, MD; Ron Munger, PhD, MPH; Maria Norton, PhD; Chiadi Onyike, MD, MHS; Truls Ostbye, MD, PhD, MPH; Ron Petersen, MD; Kathy Piercy, PhD; Carl Pieper, DrPH; Brenda Plassman, PhD; Peter Rabins, MD; Pritham Raj, MD; Russell Ray, MS; Linda Sanders, MPH; Ingmar Skoog, MD; David Steffens, MD, MHS; Martin Steinberg, MD; Marty Toohill, PhD; Leslie Toone, MS; Jeannette Townsend, MD; JoAnn Tschanz, PhD; Lauren Warren; Michael Williams, MD; and Bonita Wyse, PhD.

Back to Article Outline

References 

1. Morris CD, Carson S. Routine vitamin supplementation to prevent cardiovascular disease: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med. 2003;139(1):56–70
   • View In Article
2. Losonczy KG, Harris TB, Havlik RJ. Vitamin E and vitamin C supplement use and risk of all-cause and coronary heart disease mortality in older persons: the Established Populations for Epidemiologic Studies of the Elderly. Am J Clin Nutr. 1996;64(2):190–196
   • View In Article
   • MEDLINE
3. Lonn E, Bosch J, Yusuf S, et al. Effects of long-term vitamin E supplementation on cardiovascular events and cancer: a randomized controlled trial. JAMA. 2005;293(11):1338–1347
   • View In Article
   • CrossRef
4. Miller ER, Pastor-Barriuso R, Dalal D, et al. Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Ann Intern Med. 2005;142(1):37–46
   • View In Article
5. Breitner JC, Wyse BW, Anthony JC, et al. APOE-epsilon4 count predicts age when prevalence of AD increases, then declines: the Cache County Study. Neurology. 1999;53(2):321–331
   • View In Article
   • MEDLINE
   • CrossRef
6. Tschanz JT, Corcoran C, Skoog I, et al. Dementia: the leading predictor of death in a defined elderly population: the Cache County Study. Neurology. 2004;62(7):1156–1162
   • View In Article
   • CrossRef
7. Kerber RA, O’Brien E, Smith KR, Cawthon RM. Familial excess longevity in Utah genealogies. J Gerontol A Biol Sci Med Sci. 2001;56(3):B130–B139
   • View In Article
   • MEDLINE
   • CrossRef
8. Lindberg M, Rexstad E. Capture-recapture sampling designs. In:  El-Shaarawi AH,  Piergorsch WW editor. Encyclopedia of Environmetrics. Vol 1:Chichester: John Wiley & Sons, Ltd; 2002;p. 251–262
   • View In Article
9. Cox DR. Regression models and life-tables. J R Stat Soc. Ser B (Methodological). 1972;32(2):187–220
   • View In Article
10. Bruno RS, Leonard SW, Atkinson J, et al. Faster plasma vitamin E disappearance in smokers is normalized by vitamin C supplementation. Free Radic Biol Med. 2006;40(4):689–697
    • View In Article
    • MEDLINE
    • CrossRef
11. MRC/BHF Heart Protection Study of antioxidant vitamin supplementation in 20,536 high-risk individuals: a randomised placebo-controlled trial. Lancet. 2002;360(9326):23–33
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (422 KB)
    • CrossRef
12. Hercberg S, Galan P, Preziosi P, et al. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med. 2004;164(21):2335–2342
    • View In Article
    • MEDLINE
    • CrossRef
13. Lee IM, Cook NR, Gaziano JM, et al. Vitamin E in the primary prevention of cardiovascular disease and cancer: the Women’s Health Study: a randomized controlled trial. JAMA. 2005;294(1):56–65
    • View In Article
    • CrossRef
14. Rapola JM, Virtamo J, Ripatti S, et al. Randomised trial of alpha-tocopherol and beta-carotene supplements on incidence of major coronary events in men with previous myocardial infarction. Lancet. 1997;349(9067):1715–1720
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (64 KB)
    • CrossRef
15. 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(9004):781–786
    • View In Article
    • Abstract
    • CrossRef
16. Yusuf S, Dagenais G, Pogue J, et al. The Heart Outcomes Prevention Evaluation Study Investigators Vitamin E supplementation and cardiovascular events in high-risk patients. N Engl J Med. 2000;342(3):154–160
    • View In Article
    • MEDLINE
    • CrossRef
17. Corrigan JJ, Marcus FI. Coagulopathy associated with vitamin E ingestion. JAMA. 1974;230(9):1300–1301
    • View In Article
    • MEDLINE
18. Olson RE. The function and metabolism of vitamin K. Annu Rev Nutr. 1984;4:281–337
    • View In Article
    • MEDLINE
    • CrossRef
19. Watanabe H, Kakihana M, Ohtsuka S, Sugishita Y. Randomized, double-blind, placebo-controlled study of supplemental vitamin E on attenuation of the development of nitrate tolerance. Circulation. 1997;96(8):2545–2550
    • View In Article
    • MEDLINE
20. Munzel T, Sayegh H, Freeman BA, et al. Evidence for enhanced vascular superoxide anion production in nitrate tolerance (A novel mechanism underlying tolerance and cross-tolerance). J Clin Invest. 1995;95(1):187–194
    • View In Article
    • MEDLINE
    • CrossRef
21. Thatcher GR, Nicolescu AC, Bennett BM, Toader V. Nitrates and NO release: contemporary aspects in biological and medicinal chemistry. Free Radic Biol Med. 2004;37(8):1122–1143
    • View In Article
    • MEDLINE
    • CrossRef
22. Bowry VW, Stocker R. Tocopherol-mediated peroxidation (The prooxidant effect of vitamin E on the radical-initiated oxidation of human low-density lipoprotein). J Am Chem Soc. 1993;115(14):6029–6044
    • View In Article
    • CrossRef