“Dietary Supplements and Mortality Rate in Older Women: The Iowa Women’s Health Study” at first appears to be a large and well-laid out study. As the study describes its own design, “We assessed the use of vitamin and mineral supplements in relation to total mortality in 38,772 older women in the Iowa Women’s Health Study; mean age was 61.6 years at baseline in 1986. Supplement use was self-reported in 1986, 1997, and 2004.” It reports that a number of nutrients supposedly are linked to greater rates of death in vitamin users than in non-users. There was an associated increased risk of total mortality when compared with corresponding non-use for multivitamins, vitamin B6, folic acid, iron, magnesium, zinc and copper. Indeed, of 15 supplements analyzed by the researchers, only calcium was associated with a lower risk of mortality.
How reliable is this study and its conclusions? Not very. Many have criticized the publication vigorously as being highly misleading for reasons easy to grasp. Here are some of the concerns that have been expressed:
- This was only an observational study based on self-reporting via questionnaires and not a study establishing cause and effect.
- No attempt was made to assess the accuracy of self-reporting. As is well known from highly controlled clinical trials, many individuals will recall and report usage when they in fact are recalling and reporting their intentions. The authors cannot say whether the recall statistics were 90 percent accurate or only 50 percent accurate or varied by the supplement being taken.
- No attempt was made to independently assess the relative health of the participants or link such information to the vitamin usage. No attempt was made and there is no information on the actual nutrient status of the participants.
- Those who took vitamins at the beginning of the study reported being significantly healthier than those who did not, i.e., less hypertension, less diabetes, etc. This means that comparing those taking vitamins and those not taking vitamins was not comparing apples to apples, but potentially comparing apples to oranges because of a host of traits and conditions that might have separated these groups from the start.
- To assess the impact of vitamin usage, statistical modeling was used to remove factors that supposedly might skew the results. This modeling is highly controversial.
The statistical manipulation of the data is the worst offense of the authors of this study and it primarily is through this manipulation that they reached their negative conclusions. The authors adjusted all the numbers for “age, educational level, place of residence, diabetes mellitus, high blood pressure, body mass index (calculated as weight in kilograms divided by height in meters squared), waist to hip ratio, hormone replacement therapy, physical activity, smoking status, and intake of energy.” The results were further manipulated via yet more “multivariable adjustment” to take into account alcohol, saturated fatty acids, whole grain products, fruits, and vegetables.
It is important to understand the impact of these adjustments. Bear in mind that the women being surveyed who took supplements before the start of the study were healthier than those who did not. Jacob Teitelbaum, MD, a holistic physician and coauthor of Real Cause, Real Cure (Rodale, 2011), has provided an excellent analogy of how the authors of the Archives of Internal Medicine study biased the results via their statistical adjustments: “One could come to the same conclusion about exercise not being helpful using this same statistical approach. Take, for example a study with two control groups: those who exercise and those who do not, but both of which are put on a new exercise program. At the beginning of the study, those who already are exercisers are healthier than those who are not. By canceling this fact out, and negating the prior health of the exercisers, you can find that exercise is not beneficial.” Teitelbaum’s point is that the better health of the already exercising group may not be greatly improved by merely changing the exercise program, whereas the health of the non-exercisers will be enhanced. If the benefits of the prior exercise of the first group are subtracted from the results and then the two groups are compared, it will make prior exercise appear to be bad for health inasmuch as the prior exercise group’s health will not have improved during the course of the trial compared to the newly exercising group.
So the Archives of Internal Medicine researchers “controlled” for good health habits and for good health at the start of the observational period. Their adjustments subtracted the starting health status of the vitamin users just as in Teitelbaum’s exercise analogy. As Teitelbaum astutely observes, “The hypothesis wasn’t to state whether supplements will harm or help. It was to see if they would harm, which gives you an idea of what their study was about.”
How powerful are these adjustments, i.e., statistical manipulations? Very. Consider the actual raw findings reported in Table 2 of the study:
- Vitamin B complex was associated with a 7 percent reduction in mortality
- Vitamin C was associated with a 4 percent reduction in mortality
- Vitamin D was associated with an 8 percent reduction in mortality
- Magnesium was associated with a 3 percent reduction in mortality
- Selenium was associated with a 3 percent reduction in mortality
- Zinc was associated with a 3 percent reduction in mortality
This is not to say that all nutrients were good—for instance, copper was associated with a 31 percent increase in mortality and this negative finding well may be accurate because of the known dangers of copper overload—but the statistical adjustments significantly reversed most of the findings of the non-adjusted data. This should give one serious pause in accepting the study and its conclusions.
In short, this study would appear to be an unreliable guide to the benefits of supplements, especially multivitamin and mineral formulas. A large number of published studies, in fact, have found just the opposite, to wit, multivitamin/ mineral intake and the intake of a number of individual nutrients is associated with improved health. Here are the conclusions of just one of these studies, a study much more rigorous than that just examined: “These results indicate that use of supplements significantly improved the status of several vitamins in elderly people. Due to age-related problems concerning the intake and digestion of nutrients, a moderate, regular supplementation might be a useful option for older people who are otherwise unable to satisfy their micronutrient requirements.”
What About Vitamin E and Other Antioxidants?
Do the questionable findings of the Archives of Internal Medicine study mean that other studies with negative conclusions about supplements are wrong? Not necessarily. The JAMA study is a case in point. It found that men who took vitamin E for prostate health were 17 percent more likely to develop prostate cancer than those not supplementing. The authors of this study warned that the “observed 17 percent increase in prostate cancer incidence demonstrates the potential for seemingly innocuous yet biologically active substances such as vitamins to cause harm.” In this instance, the findings most likely are accurate. Nevertheless, there are problems in drawing conclusions about supplements from this study.
This study builds upon an earlier study published in 2008, the “Selenium and Vitamin E Cancer Prevention Trial,” which found a “statistically non-significant” increase in prostate cancer risk in those taking 400 IU of vitamin E daily. In the earlier study, the primary analysis included 34,887 men who were randomly assigned to 1 of 4 treatment groups: 8,752 to receive selenium; 8,737, vitamin E; 8,702, both agents, and 8,696, placebo. The 2011 trial included 54,464 additional person-years of follow-up and 521 additional cases of prostate cancer since the primary report. Compared with placebo, the absolute increase in risk of prostate cancer per 1000 person-years was 1.6 for vitamin E, 0.8 for selenium, and 0.4 for the combination. This led to the conclusion: “Dietary supplementation with vitamin E significantly increased the risk of prostate cancer among healthy men.”
The Council for Responsible Nutrition (CRN) and other observers have admitted that this JAMA study was well done—and, indeed, it was well performed within its stated parameters. The issue is the degree to which its results can be generalized. The form of vitamin E studied was the synthetic alpha-tocopherol. This likely is the real reason that the study failed. In several other studies on control of inflammation, prevention of cancer, etc., gamma-tocopherol and deltatocopherol have exhibited benefits, whereas alpha-tocopherol has not. In fact, the excessive intake of alpha-tocopherol long has been known to be potentially harmful. Here are some reasons why:
- Supplementation with alpha-tocopherol clearly depresses gamma-tocopherol levels, whereas gamma-tocopherol has little or no impact on alpha-tocopherol levels. Supplementation with alpha-tocopherol clearly depresses delta-tocopherol levels, whereas delta-tocopherol has little or no impact on alpha-tocopherol levels.
- Alpha-Tocopherol also attenuates the benefits of gamma-tocotrienol on cardiovascular disease and at high levels may interfere with the benefits of the four tocotrienols more generally.
- High chronic intake of alpha-tocopherol from any source may compromise mitochondrial ubiquinol (CoQ10) metabolism. Excessive oxidative degradation of tocopherol can potentially interfere with mitochondrial electron transfer.
The tocotrienols family of vitamin E and gamma- and delta-tocopherols all seem to offer only positive results. To the contrary, alpha-tocopherol supplementation, regardless whether the source is natural or otherwise, at this point in time appears to be questionable.
Vitamin E is an exemplar of a more general trend in the research regarding antioxidants. Because it is the premier example of an antioxidant that breaks free radical chain reactions, it initially was expected to have positive results against cancer, heart disease, and so forth and so on. For alpha-tocopherol after numerous large and lengthy trials, this promise no longer is supportable. The era of excessive expectations for oral antioxidants is being tamed by the realism of clinical and other research. Inasmuch as expectations for supplements in general were bolstered by the antioxidant theory and the evidence apparently supporting it, the failures of large extended clinical trials to unequivocally support vitamin E and, to a lesser extent, additional antioxidants, such as selenium, cannot but damage the image of other supplements. Today, it has become apparent that many of the benefits of the tocotrienols, gammatocopherol, etc., clearly are not due to their antioxidant and free radical trapping and chain-breaking functions and, instead, are due to other mechanisms of action.
Vitamin E (as alpha-tocopherol), selenium and beta-carotene are three antioxidants whose stars have been tarnished the most in recent years. Under certain circumstances, excessive intakes of each of these have proven to be potentially harmful. It should be noted that two of these are fat-soluble, hence can build up in the body, and the third is a trace element. In contrast, no realistic general concerns have been demonstrated with regard to vitamin C, which is water soluble and easily eliminated from the body.
Balanced and reasonably dosed multivitamin and mineral supplements have proven track records not only for safety, but also for insuring nutritional adequacy in groups, such as the elderly and teenagers, who otherwise often experience significant shortfalls in essential nutrient intakes. Most studies on multivitamin and mineral supplements have found that these contribute to improved health. Nevertheless, expectations should be reasonable.
Very high intakes of individual supplements being taken essentially as drugs is another issue entirely. Consuming daily the amount of alpha-tocopherol found in several bushels of wheat or the amount of catechins found in 30 cups of green tea is a practice that always needs to be examined critically. With some nutrients, there are significant benefits with little possibility of harm. However, as the case of alpha-tocopherol demonstrates, this is not always the case.