Two recently published scientific articles have been picked up by the mainstream press and are being used to cast a negative light on dietary supplements. We would like to briefly explain the two studies and help to put their findings in context.
Multivitamins
Mursu J, Robien K, Harnack LJ, Park K, Jacobs DR. Dietary supplements and mortality rate in older women. Arch Intern Med. 2011;171(18):1625-33.
The first of the two articles was based on an analysis of data from the Iowa Women’s Health Study. This study, started in 1986, was designed to investigate associations between diet and lifestyle factors and the incidence of cancer in postmenopausal women. The current analysis examined the connection between supplement use and the risk of mortality in a cohort of over 38,000 women enrolled in the Iowa Women’s Health Study. The authors conclude that common vitamin and mineral supplementation by older women is of no benefit and is associated with an increased total mortality risk. The authors point out that the connection between supplementation and increased mortality risk was strongest for iron; however, they also found that supplemental multivitamins, folic acid, B6, magnesium, and zinc were associated with an increased risk of mortality. This has been portrayed by the media as new findings of the “deadly danger” of vitamins, and a further opportunity for those that believe the industry needs more regulation to advance their agenda. So, let’s put some things in context and better understand some key points about this study:
• This is an observational study spanning a 22 year time frame (1986-2008) and is not an intervention trial in which supplements were given. Observational studies are certainly not intended to determine cause and effect relationships.
• The use of supplements was obtained by subjects self-reporting their use at three times during the course of the study – 1986 (baseline), 1997, and 2004. In the authors’ analysis of mortality risk between the years of 1986 and 2004, if a subject reported supplement use during any of the 3 time points, they were considered to be a supplement user for the entire duration of the study. Certainly this methodology may not reflect long-term supplement use, as subjects could have discontinued or started supplementation that would not have been reflected in the analysis. Additionally, as the authors noted, supplement use changed dramatically during the course of the study. In 1986, self-reported supplement use was 62.7%, whereas in 2004 supplement use was 85.1%. Increased supplement use during the course of the study could have been the result of many factors, including self selection by subjects attempting to deal with failing health.
• The authors had to manipulate and adjust their analysis in order to find the negative associations that they reported. Supplement use at baseline was associated with a lower prevalence of diabetes, hypertension, and smoking as well as a lower BMI and hip:waist ratio. Supplement users were also more likely to have a lower intake of total calories, fat, and saturated fat and had a greater intake of grains, fruits and vegetable. When the data were analyzed and adjusted for age and energy intake, many of the supplements assessed were associated with a significantly decreased risk of mortality. It was not until the data were adjusted in an attempt to control for the variables associated with supplement use that the negative associations became statistically significant. However, in a statistical model it may be impossible to fully separate the healthful behaviors and characteristics that are associated with being a supplement user.
• The increased mortality risk from iron supplementation was mostly observed in subjects supplementing with greater than 200 mg/day. This level of supplementation is often associated with conditions such as anemia, for which there may already be an increased risk of mortality. This was not controlled for in the current study.
• The use of hormone replacement therapy was significantly greater in supplement users (13.5%) compared with nonusers (7.2%) during the duration of the study. While the authors attempted to adjust for this in their analyses, it certainly could have represented a confounding variable that could not have been sufficiently factored out of the analysis.
• Importantly, the manner in which the authors classified “nonusers” of supplements in their analyses is curious. They looked at 15 different supplement variables: multivitamins, vitamin A, beta-carotene, B6, folic acid, B complex, vitamin C, vitamin D, vitamin E, calcium, copper, iron, magnesium, selenium, and zinc. To assess the risk from a particular variable such as zinc, they compared the mortality in users of zinc supplements with the mortality in “nonusers” of zinc supplements. However, while the subjects in the “nonusers” of zinc category may not have been taking a standalone zinc product, many would have also been taking a multivitamin that contained zinc or supplements other than a standalone zinc product. As a result, in the analysis of risk for zinc users, the group against which they were being compared would have contained subjects who were consuming supplemental zinc. This point is not addressed in the study and could represent a serious limitation with respect to the interpretation of the data.
• It is interesting to note that very little attention is given to the fact that calcium supplementation was associated with a significant reduction in total mortality risk, cardiovascular mortality risk and cancer mortality risk.
These points highlight just a few of the questions and limitations that surround this study. The commentary accompanying the study is equally biased against supplementation and in general portrays supplementation in a negative light. We don’t believe that there is anything contained within this study that should cause one to stop taking their supplements, and in fact it raises more questions than it answers.
Vitamin E
The second recently published article was an analysis of data from the SELECT trial.
Klein EA, Thompson IM, Tangen CT, et al. Vitamin E and the risk of prostate cancer. The Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA. 2011;306(14):1549-1556.
Similar to the multivitamin trial, the SELECT trial received significant negative media attention due to the reported finding of an increased risk of prostate cancer in those supplementing with vitamin E. The SELECT trial was a prospective randomized, placebo-controlled trial investigating if vitamin E, selenium, or the combination could reduce the risk of prostate cancer. Over 35,000 men at average risk for prostate cancer were randomized between 2001 and 2004 to participate in the study. Subjects were assigned to 400 IU/day dl-alpha-tocopherol (synthetic alpha-tocopherol), 200 μg selenium (from selenomethionine), the combination of the two, or placebo. A 2009 publication from the SELECT trial, with a median follow up of 5.5 years, revealed no statistically significant increase or decrease in prostate cancer risk among any of the groups. In 2008, the data and safety monitoring committee recommended that the study be discontinued because of “lack of efficacy for risk reduction…” The current publication is an analysis of the data taking in to account additional cases of prostate cancer in subjects through May, 2011. The authors reported a 17% increase in the risk of prostate cancer in the vitamin E group and did not find an increased risk in either the selenium or selenium plus vitamin E groups. A number of points should be understood in order to put these findings in context.
• The results of this study are in contrast to previously published randomized clinical trials including the Alpha-Tocopherol, Beta-Carotene trial (ATBC) and the Physicians Health Study II, which either found a reduced risk of prostate cancer or no change in risk.
• The results are also in contrast to previous trial including the CHAOS, HOPE, HOPE TOO and SPACE trials.
• It should be noted that the expert panel assembled by the American Institute for Cancer Research/World Cancer Research Fund (2007) examined a total of six cohort studies, 14 case-control studies, and one ecological study regarding the relationship between dietary and/or serum vitamin E and concluded that “there is limited evidence suggesting that foods containing vitamin E, and alpha-tocopherol supplements are protective” against the risk of prostate cancer. Importantly, none of these observational studies reported an increased risk of prostate cancer associated with vitamin E.
• A strange finding of the current study is that vitamin E was very specific for prostate cancer risk. No other increased risk of any other cancer, cardiovascular disease, diabetes or all-cause mortality was observed in the vitamin E group. The authors were not able to offer, nor does there appear to be any biological reasoning for the apparent specificity on prostate cancer risk. The authors also could not explain why there was no increased from the presence of selenium with vitamin E.
• No data on vitamin E levels were given so we don’t know the baseline vitamin E status of the subjects in any of the groups.
• Although the study was stopped in 2008, the current analysis includes cancers through 2011. We don’t know how many subjects continued taking supplements on their own and what impact the lack of supplementation over the past three years could have had on the outcome.
• Prostate cancer status was determined by self-reporting at 6-month intervals. The diagnosis was then centrally confirmed by the SELECT central pathology laboratory after obtaining the subject’s medical records. However, the authors state that 17% of the cancers were not centrally confirmed yet are still included in the analysis.
• The SELECT trial had an unusually small number of deaths from prostate cancer compared with what would be expected given the length of the study and number of participants.
• The form of vitamin E used was synthetic alpha-tocopherol and not natural-source alpha-tocopherol. We know that these forms of alpha-tocopherol are significantly different and are metabolized differently by the body.
As is the general rule, one should not rush to judgment based on the result of a single study, especially when it does not fit with the totality of the existing data. All of the evidence should be considered when making decisions about supplementation. Similar to the multivitamin study, many important questions remain unanswered in the most recent SELECT analysis.
