Dietary supplements and mortality in older women – a response to a scientific publication

Hohenheim Nutrition Information, Institute of Biological Chemistry and Nutrition Science, Hohenheim University, Germany

“In 2009, Kyong Park and colleagues published a study of more than 40,000 women taking part in the Iowa Women's Health Study, whom they had interviewed over a period of 19 years with questionnaires on the consumption of vitamin and mineral supplements (1). They came to the conclusion that the use of dietary supplements among older people was especially positive because reduced nutrient intake, likely as a result of reduced food intake in old age, could be compensated.

So it was all the more surprising that the same authors came to an astonishing new conclusion with a recently published re-evaluation of the study participants’ data. In this study (2) the authors conclude that older women who take dietary supplements have a higher risk of death. This, according to the authors, is especially the case with iron supplements and multivitamin/mineral preparations. At the same time, the authors saw a reduced risk of death in women who took calcium supplements. The study period extended over 22 years and, at the beginning (1986), covered 38,772 middle-age women of 61.5 years, 24,329 of whom reported using a dietary supplement – without stating the exact dose or the reasons for taking it (preventative, therapeutic, etc.) – and 14,443 of whom said they did not. By the closing date of 31.12.2008, a total of 15,594 (40.2%) of the women had died.

According to the authors, the consumption of multivitamin/mineral preparations increased the mortality rate among older women by 2.4%, supplements containing copper increased mortality by 18%, and calcium supplementation reduced mortality by 3.8%. For all other dietary supplements, even iron, there were no statistically significant differences shown in the end.

How can the observed or described increase in mortality be explained?

• The study investigates the intake of dietary supplements once in 1986, then again in 1997 and most recently in 2004. The authors acknowledge that the questionnaire used for this purpose was not validated for supplement intake. This certainly does not fulfill the scientific conventions for collecting such data.

• During the survey, the duration of intake and the possible changes in the supplement intake were barely recorded; the reason for intake was also not provided. It is entirely possible that intake could have taken place in the context of a disease, which would have made a significant contribution to a shortening of life. Iron supplements, for example, are prescribed for the treatment of iron deficiency, i.e. when anemia is detected. Anemia is often the result or a side effect of an illness that can in turn be linked to a higher risk of death compared with healthy people.

• It is also completely unclear whether the women who stated that they took dietary supplements in 1986 still took them in 1997 or 2004. Nor was it clear whether they had changed supplements or had stopped using them since. The same applies the other way around for women who stated that they took no supplements in 1986. The authors note that a major weakness in the study lay in the low number of responses to the survey in 1986 and to the follow-up survey. Thus, the possibility exists that non-response to questionnaires is already a selective assessment per se.

• Next, the authors say that women who stated that they took B-complex vitamins and vitamins C, D, E and calcium at the beginning of the study showed a significantly lower risk of death than those who took no supplements. Copper, on the other hand, is supposed to be associated with a higher risk. When these data were further ‘adjusted,’ i.e., with regard to various lifestyle factors such as smoking, alcohol consumption, etc., the lower risk of death disappeared, which left only the significant result obtained for calcium.

It turns out that it’s possible to influence the results as desired by using different adjustment methods. The authors provide the best evidence for this: by arbitrarily choosing which factors to consider, these could ultimately reduce or increase the risk of death for individual dietary supplements. For those women who took B-complex vitamins, as well as vitamins C, E and D and calcium, mortality was at first significantly lower. After adjustment of the data, the significance largely disappeared, remaining only for calcium. Only after a further adjustment was the mortality increase result obtained.

Overall, there are no scientifically accurate data to be gleaned in this study that could support the conclusion of the study that the intake of dietary supplements shortens life.

On the basis of various studies, which are also mentioned by the authors, it is apparent that in the case of older persons in particular, quality of life increases with the intake of multivitamin/mineral supplements or various dietary supplements (especially vitamin D), and that these have little influence on mortality. That said, it is just the study itself, driven by unclear motivations, that must be scrutinized. We should ask ourselves to what extent the conclusion, which is unhelpful for older people (among whom a high percentage of deficiency has been detected), is leading us astray.

The same applies to a comment (3) explaining that antioxidants such as those found in dietary supplements were known to increase mortality. As evidence, it cites three studies that only suggested this for long-term use of high-dose supplements (of ten times or more of the recommended daily dose). The commentator also notes that the use of vitamin and mineral supplements cannot be recommended for a population already sufficiently supplied with micronutrients. This is a scientific consensus. However, the group covered in the current study is composed of older participants and it can be assumed, based on a variety of data, that the declining food intake of older people no longer meets qualitative requirements (micronutrient content).

It must be maintained that the results of the current study (2) are implausible and unhelpful for discussing the problems of malnutrition because the terse statement that dietary supplements are dangerous is itself not very helpful and may alarm those whom it potentially concerns. After critical examination of the work the uncomfortable feeling remains that the authors, much like the journal that took on the publication of this work, wanted to exploit a negative message to increase the importance of the study.”

Hohenheim, October 2011