Written by Neil E. Levin, CCN, DANLA, Nutrition Education Manager.  The meta-analysis published in JAMA that reported antioxidant vitamins do not extend life and even increase death rates slightly had many variables, flaws, and an arbitrary structuring of its statistical model. 

A meta-analysis reported in the medical journal JAMA this week reported that antioxidant vitamins do not extend life and may even increase death rates slightly. (1) These conclusions make no sense, based on the scientific record. A meta-analysis relies on a statistical model of existing science, and this model has severe limitations.  Even the authors admit some of the basic problems inherent in this type of analysis.

More importantly, the authors could not find a dose-dependent or cause-and-effect relationship between antioxidants and deaths (from all causes) of study participants.  In other words, they couldn’t show that antioxidants actually caused any deaths or that there was risk at a particular dosage.  Yet this questionable speculation received widespread publicity from the sensation-hungry media during “Sweeps Month”, dutifully spreading the lie that antioxidants are now worthless and dangerous.

The researchers pooled 68 previously published trials but arbitrarily excluded all published studies that had no deaths reported from any cause.  Indeed, 405 otherwise eligible studies were excluded solely for this reason, which if included would likely have dramatically changed the results and conclusion.  The researchers did not disclose why they decided to exclude these.  This is equivalent to playing a card game after removing all but 7 cards from the deck. (That wouldn’t be a fair game, would it?) They largely ignored the original outcome measures of the studies, many of which had shown positive results for antioxidants, to look only for deaths from any cause in a tiny segment of all published research.

This arbitrary decision echoes a frequently cited complaint by scientists commenting to the journal Annals of Internal Medicine when the infamous Miller meta-analysis of vitamin E was released a few years ago, which led to a dramatic slowdown of vitamin E sales. (2) It is interesting that the Miller study’s negative conclusions about vitamin E safety have since been thoroughly debunked by a more rigorous analysis published in the American Journal of Clinical Nutrition by leading antioxidant experts. (3) It is even more interesting that the flawed Miller review was cited as a reference by the JAMA authors but the second, more thorough analysis of the same data by real nutrition experts was not.  The lesson learned is that a flawed meta-analysis of nutrients by statisticians and physicians may not hold up to a more competent review done by actual experts in the field of nutrient interactions, though the initial report may have scared people and changed their behavior. Critical comments and corrections typically go ignored, uncited and unreported, in contrast with the sensational initial report. (4)

I question both the selection of studies reviewed and the references cited in this meta-analysis.

Obviously, excluding six times as many potentially eligible studies as were actually chosen solely because of a requirement that someone in the study population had to die unfairly magnifies negative results by dramatically reducing the pool of studies with potentially positive results and healthier populations. This negative shift is a result of limiting the combined patient population to those studies with at least one dying patient. This population shifts to those individuals who are more likely to be deficient in a variety of antioxidant substances and who are unlikely to respond to limited amounts of one or few supplemental antioxidants.

The lack of additional supporting antioxidants may even sometimes increase the oxidative stress on the body. A review of antioxidant science noted, “These negative results…should not be taken as evidence that the free radical theory of aging is flawed.  In fact, they prove merely that a complex organism like a human or rodent is unlikely to respond predictably to crude manipulations such as supplementation with one or a small number of compounds.” (1)(5)(6) This mirrors the JAMA authors’ admissions that “antioxidant supplements may show interdependency and may have effects only if given in combination,” and that their findings “should not be translated to potential effects of fruits or vegetables,” which are sources of numerous and varied antioxidant substances. (1)

Previous studies have shown the folly of such a protocol.  Some years ago an antioxidant study in Finland was halted early because of a widely reported increase in cancer rates among male smokers taking beta-carotene. (7) Headlines associated this supplement with cancer risk. Despite objections that the study was flawed, beta-carotene use dropped.

A later analysis published in July 2004 took another look at that same Finnish smokers’ study data, but now taking into account total antioxidant intake, which (should have) cleared away the scientific controversy. (8) A composite antioxidant index was generated for each of the 27,000 men over 14 years. The calculated amounts of carotenoids, flavonoids, Vitamin E, selenium and vitamin C were compared to actual lung cancer rates, with a clear result: an increased intake of a combination of antioxidants lowered lung cancer risk in male smokers. Another large study has noted that high carotenoid intake, as confirmed by measures of blood levels, was associated with lower mortality rates among the elderly over a ten year period. (9)

The dietary level of antioxidants is an independent predictor of plasma beta-carotene, especially in moderate alcohol drinkers. A more recent study reports, “This may explain, at least in part, the inverse relationship observed between plasma beta-carotene and risk of chronic diseases associated to high levels of oxidative stress (i.e., diabetes and CVD), as well as the failure of beta-carotene supplements alone in reducing such risk.” (10) In other words, we shouldn’t expect one or two supplemented antioxidants to compensate for a deficiency of total antioxidants in the diet.  In fact, many of the protocols for supplementation in the included studies may have actually been of too low potency to achieve noticeable health benefits by remedying latent nutrient deficiencies in fragile patient populations. In other words: many of the interventions were too little, too late. Don’t blame the vitamins.

The JAMA report admits that the study populations, the variety of antioxidants used, their potencies and the protocols for taking them were extremely variable, complicating their data with many uncompensated variables. Yet the authors actually claim that, “This increases the trustworthiness of our findings.” I don’t think so! One trial included gave only a single serving of antioxidants and then monitored participants for 3 months.  Others used doses of as little as 10 IU of vitamin E (a low amount that is below the Daily Value) and 20 mg of selenium 1 (an amount far below the 70 mcg DV and not anywhere near the 200+ mg/day associated with lower cancer rates). (11)(12)

I am not alone in these criticisms. Alexander Schauss, PhD, FACN has written, “The range of doses in the different trials they selected for the meta-analysis is dramatic. For example, vitamin A ranged from 1333 IU to 200,000 IU, and vitamin E from 10 IU to 1000 IU. The duration of the studies range from 28 days to 12 years. Nevertheless they were all lumped together.”

An Associated Press article quoted other experts criticizing this meta-analysis: ‘Meir Stampfer, professor of nutrition and epidemiology at the Harvard School of Public Health, said the new analysis hasn’t discouraged him from taking his vitamins. Stampfer said the studies were too diverse to pool together because they looked at various combinations and doses of antioxidants tested in different groups of people. The trials ranged from a three-month study of 109 elderly nursing home residents to a 12-year study of 22,071 male doctors. “This study does not advance our understanding, and could easily lead to misinterpretation of the data,” said Stampfer, who was not connected to the new report.’

The AP report also quoted Donald Berry, chairman of the department of biostatistics at the University of Texas’ M.D. Anderson Cancer Center, stating that this expert also disagreed with the researchers’ finding of an increased risk of dying. “There are so many choices you can make when you’re doing these analyses,” he said.

A study of approximately 90,000 nurses suggested that the incidence of heart disease was 30% to 40% lower among nurses with the highest intake of vitamin E from diet and supplements. Researchers found that the apparent benefit was mainly associated with intake of vitamin E from dietary supplements. High vitamin E intake from food was not associated with significant cardiac risk reduction. (13)

Levels of Vitamin E above 100 IU daily are associated with decreased risk of coronary heart disease and certain types of cancer, as well as enhancement of immune function. These increased vitamin E intakes are considerably above levels obtainable from diet alone. (14)(15)(16) In a report on the Women’s Health Study published in JAMA, subjects supplementing with vitamin E were reported to have a significant 24% reduction in cardiovascular deaths. (17)

Have these previously published benefits of antioxidants miraculously vanished simply because some doctors manipulated a statistical model to elicit unreliable data with no solid basis?

Many of the studies included were of patients with specific, serious medical conditions, including one of elderly nursing home patients measuring incidences of bacterial infections (contrasting with a 2004 study published in JAMA noted that, “we observed a protective effect of vitamin E supplementation on upper respiratory tract infections, particularly the common cold, that merits further investigation.”( 18), patients with tumors removed from their colon/rectum (antioxidants are associated with apoptosis, a desirable change that leads to death of cancer cells (19)( 20), patients with age-related macular degeneration (a condition associated with a deficiency of various antioxidants (21-24), patients with coronary heart disease (a condition related to oxidative damage (14-16), dialysis patients with a history of cardiovascular disease (a condition related to oxidative damage that is reduced by supplemental vitamin E (17), cataract patients (another condition related to oxidative damage (25), male cigarette smokers/present and former cigarette smokers/asbestos workers (all related to low levels of total antioxidants and high toxic load), as well as patients with alcoholic hepatitis, cirrhosis, lupus, heart failure, ALS, etc.

This meta-analysis will not stand the test of time because of its many variables, flaws and the arbitrary structuring of its statistical model. When better studies exist, often supported by blood assays, that show higher serum antioxidant levels reduce actual death rates in large populations, then no arbitrary statistical model should be able to negate that robust science with a merely theoretical danger based on such preliminary, questionable criteria.

All studies cited here were published in peer-reviewed scientific journals, but that does not make them all of equal quality. Remember my story of the meta-analysis on vitamin E that was refuted by a better meta-analysis, yet both were peer-reviewed? A meta-analysis has more validity if fewer variables are included and if the selection of studies included is not biased by a presumed conclusion. Were hundreds of studies without dying participants inelegible for this particular meta-analysis review simply because of a selection bias, with an intent to demonstrate the dangers of supplementation?

These scientists should know better. I see their report as an ill-disguised partisan attack by medical special interests on dietary supplements, a smokescreen for those that don’t look at the quality and quantity of well-designed studies that do show the benefits of vitamins to protect health and prevent deaths. The Lewin Group has presented evidence that the use of antioxidants could save the vision and independence of many senior citizens, while saving the public billions of dollars in healthcare costs. 26 The Institute of Medicine, part of the National Institutes of Health, has reviewed hundreds of well-designed studies and has set safe upper limits for some antioxidants at levels far above the Daily Values. (27)

Antioxidants are safe, and proven so in better studies than this one.

http://honestnutrition.blogspot.com/2007/03/antioxidant-confusion.html

REFERENCES:

  1. Bjelakovic G, et.al. Mortality in Randomized Trials of Antioxidant Supplements for Primary and Secondary Prevention: Systematic Review and Meta-analysis. JAMA 2007. 297(8):842-857
  2. Miller ER 3rd, et al. Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Ann Intern Med. 2005 Jan 4;142(1):37-46. Epub 2004 Nov 10. Summary for patients in: Ann Intern Med. 2005 Jan 4;142(1):I40. PMID: 15537682
  3. Hathcock JN, et al. Vitamins E and C are safe across a broad range of intakes. Am J Clin Nutr. 2005 Apr;81(4):736-45. Review. PMID: 15817846
  4. Levin, N. Land of Confusion: How Poor Science and Misleading Media Coverage Create Public Confusion About How Dietary Supplements Affect Health. J App Nutr, Vol 55, No. 1, 2005 8-15
  5. Beckman KB, Ames BN. The free radical theory of aging matures. Physiol Rev. 1998 Apr;78(2):547-81. Review. PMID: 9562038
  6. Block, G. Are clinical trials really the answer? Am. J. Clin. Nutr. 62, Suppl.: 15175-15205, 1995
  7. 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 Apr 14;330(15):1029-35. http://content.nejm.org/cgi/content/full/330/15/1029?ijkey=bd47b716724d0…
  8. Wright ME, et al. Development of a Comprehensive Dietary Antioxidant Index and Application to Lung Cancer Risk in a Cohort of Male Smokers. July 2004 American Journal of Epidemiology http://aje.oupjournals.org/cgi/content/abstract/160/1/68?maxtoshow=&HITS…
  9. Buijsse B, et al. Plasma carotene and alpha-tocopherol in relation to 10-y all-cause and cause-specific mortality in European elderly: The Survey in Europe on Nutrition and the Elderly, a Concerted Action (SENECA). Am J Clin Nutr 2005;82:879–886.
  10. Brighenti F. The total antioxidant capacity of the diet is an independent predictor of plasma beta-carotene. European Journal of Clinical Nutrition (2007) 61, 69–76.
  11. Clark LC, Marshall JR. Randomized, controlled chemoprevention trials in populations at very high risk for prostate cancer: elevated prostate-specific antigen and high-grade prostatic intraepithelial neoplasia, Urology 57 (2001), pp. 185–187.
  12. Duffield-Lillico, AJ, et al. Baseline characteristics and the effect of selenium supplementation on cancer incidence in a randomized clinical trial: a summary report of the Nutritional Prevention of Cancer Trial, Cancer Epidemiol. Biomarkers Prev. 11 (2002), pp. 630–639.
  13. Stampfer MJ, et al. Vitamin E consumption and the risk of coronary disease in women. N Engl J Med 1993;328:1444-9
  14. Bauernfeind, J. Tocopherols in Foods. In: Vitamin E: A Comprehensive Treatise. Marcel Dekker, Inc., New York and Basel, pp. 99-167, 1980.
  15. Horwitt, M.K. The Promotion of Vitamin E. J. Nutr. 116:1371-1377, 1986.
  16. Weber, P., Bendich, A. and Machlin, L.J. Vitamin E and Human Health: Rationale for Determining Recommended Intake Levels. Nutrition 13:450-460, 1997.
  17. I-Min Lee, MBBS, ScD; et al. Vitamin E in the Primary Prevention of Cardiovascular Disease and Cancer. The Women’s Health Study: A Randomized Controlled Trial. JAMA. 2005;294:56-65
  18. Meydani SN, et al. Vitamin E and respiratory tract infections in elderly nursing home residents: a randomized controlled trial. JAMA. 2004 Aug 18;292(7):828-36. Erratum in: JAMA. 2004 Sep 15;292(11):1305. PMID: 15315997
  19. Narayanan BA. Chemopreventive agents alters global gene expression pattern: predicting their mode of action and targets. Curr Cancer Drug Targets. 2006 Dec;6(8):711-27. Review. PMID: 17168675
  20. Valko M, et al. Free radicals and antioxidants in normal physiological functions and human disease. Int J Biochem Cell Biol. 2007;39(1):44-84. Epub 2006 Aug 4. Review. PMID: 16978905
  21. Chiu CJ, Taylor A. Nutritional antioxidants and age-related cataract and maculopathy. Exp Eye Res. 2007 Feb;84(2):229-45. Epub 2006 Jul 31. Review. PMID: 16879819
  22. Moriarty-Craige SE, et al. Antioxidant supplements prevent oxidation of cysteine/cystine redox in patients with age-related macular degeneration. Am J Ophthalmol. 2005 Dec;140(6):1020-6. PMID: 16376645
  23. Richer S, et al. Double-masked, placebo-controlled, randomized trial of lutein and antioxidant supplementation in the intervention of atrophic age-related macular degeneration: the Veterans LAST study (Lutein Antioxidant Supplementation Trial). Optometry. 2004 Apr;75(4):216-30. PMID: 15117055
  24. Koh HH, et al. Macular Pigment Optical Density in Early, Age-Related Maculopathy (ARM); Comparisons With Normals and Effects of a Lutein Supplement. Invest Ophthalmol Vis Sci 2002;43:2562
  25. Meyer CH, Sekundo W. Nutritional supplementation to prevent cataract formation. Dev Ophthalmol. 2005;38:103-19. Review. PMID:  15604620
  26. Vanzo JE, et al. An Evidence-Based Study of the Role of Dietary Supplements in Helping Seniors Maintain their Independence. The Lewin Group Inc. January 20, 2006
  27. National Institutes of Health, Institute of Medicine, Office of Dietary Supplements. Vitamin E Fact Sheet.