The use of dietary supplements is among the most intensely discussed topics in contemporary nutrition. Results of related studies are regularly highlighted in the lay press, sometimes generating more confusion than clarity among the general public. In any case, it seems that most people have not waited for the scientific evidence to come in to make their personal decisions about using supplements: more than half the US adult population are already taking some form of dietary supplements.
In that context, it was very timely that the National Institutes of Health decided to convene a "state-of-the-science" conference to address the issue of vitamin/mineral supplements. This type of conference, unlike the "consensus" conferences also hosted by the National Institutes of Health, does not aim to issue a consensus recommendation for health professionals and the public. Instead, the aim of the state-of-the-science conference is precisely that, to summarize our current knowledge on the topic and to identify gaps that require further research. This format seemed the best suited to define the current knowledge platform and to separate fact from fiction on a highly controversial subject.
In preparation for the conference, the National Institutes of Health commissioned a comprehensive, systematic review of the literature on the topic of vitamin-mineral supplements. The task was carried out by a team at the Evidence-Based Practice Center at Johns Hopkins University, led by Dr Han-Yao Huang.1
The review was based on a literature search of MEDLINE, EMBASE, and the Cochrane database for the period of January 1966 to February 2006, targeting selected vitamins and minerals, with specific disease end points. Multinutrient was defined as supplements containing at least 3 vitamins and/or minerals, excluding formulations containing drugs, hormones, or herbs. Supplements containing 2 related vitamin/minerals (eg, calcium and vitamin D, folic acid and B12) were considered as single-nutrient supplements. Only randomized controlled trials (RCTs) in adult populations or previous systematic reviews were included. Each article underwent a sequential review by 2 or 3 separate reviewers that included title, abstract, and a final inclusion or exclusion review decision. To address safety, the selection criteria were expanded to include studies in children, longitudinal studies, and case reports. The full report is available online at http://www.ahrq.gov/downloads/pub/evidence/pdf/multivit/multivit.pdf.
Of the 11,324 publications in the initial search, only 63 met the selection criteria and were included in the final review. For single-nutrient supplements, 6 RCTs for beta-carotene (22 publications), 4 for vitamin E (16 publications), and 2 for selenium (6 publications) were included. For multivitamin/minerals, there were 5 RCTs reported in 10 publications. The remainder were articles selected because of their safety data.
An initial and remarkable finding of this review was the scarcity of RCTs on the efficacy of vitamin/mineral supplements for disease prevention. The vast majority of the literature captured by the initial search ranged from a few good cross-sectional or observational studies to many small, uncontrolled, poorly documented studies with inadequate exposures to assess effects. Furthermore, many studies did not report disease outcomes but only intermediary end points, such as blood cholesterol or antioxidant levels. Another interesting finding was that almost no study used supplements comparable with commercially available multivitamin/mineral products. Instead, a wide range of doses and combinations were used, sometimes differing by several-fold. Thus, results from these trials may not be always applicable to consumers' choices of commercially available supplement products.
Overall, the data from RCTs did not demonstrate a consistent preventive effect of supplementation on cancer, cardiovascular disease, diabetes, or eye disease (cataracts or macular degeneration). No eligible studies were found for many of the other targeted diseases. Still, some trials did report significant effects. For example, the French SUVIMAX study, which administered a combination of vitamins C and E, beta-carotene, selenium, and zinc, reported a 31% reduction in overall cancer rates in men but not in women. The AREDS study found a beneficial effect of zinc, alone or with other antioxidants, in slowing the progression of age-related macular degeneration, more consistently for the moderate stage of the disease.
The evidence base for assessing safety was also quite limited because even well-designed RCTs did not always collect adequate safety data. The final review of safety included 10 studies using multivitamin/mineral supplements and 24 using single-nutrient supplements. Important limitations of the safety data included the following: (a) adverse events were self-reported; (b) events were reported only for participants who dropped out of the study; and (c) criteria and scales for defining adverse events and grading severity were nonstandard and/or subjective. Overall, there was no consistent pattern of increased risk associated with the use of supplements. Nevertheless, some side effects were repeatedly reported for some supplements, for example, yellowing of the skin after regular consumption of beta-carotene (alone or in a multivitamin formula). One study reported increased incidence of kidney stones in women using 1 g/d of calcium carbonate, and 2 studies found that vitamin A supplementation for 1 year or more increased serum triglyceride levels. Also clear, and more alarming, was the increased rates of lung cancer associated with beta-carotene supplementation in smokers or in persons exposed to asbestos.
Where do we stand after this important effort to clarify the use of vitamin/mineral supplements? First, the remarkably small number of well-designed, credible efficacy studies is sobering. Second, the heterogeneity of supplement formulations used in different studies makes it difficult to draw a general conclusion applicable to consumers of commercial multivitamin/mineral products. Third, although overall there was no consistent evidence of adverse effects, this was due in part to the fact that many trials did not actively collect safety information or did so in a nonsystematic and/or limited manner.
Recently completed dietary supplement trials have highlighted the difficulties in performing clean, rigorous supplement trials in contemporary US society. It is becoming increasingly difficult to identify groups that have never used dietary supplements or even those who are willing to stop taking them for several years when enrolled in a study. Investigators may have no choice but to allow participants to continue to use "personal" supplements during the study, as was done in some recent trials. Thus, it may not be realistic to expect that several new, large-scale, definitive trials will soon clarify the outstanding questions on supplement use. In the meantime, it may be opportune to remind us that the 2005 Dietary Guidelines for Americans reported that the US adult population consume inadequate amounts of 7 key nutrients: calcium, vitamin A/carotenoids, vitamins C and E, fiber, magnesium, and potassium.2 We also know that the American diet has too much saturated fat and sodium and too many calories. Clearly, there is plenty of room for improvement, and virtually all of it (plus more physical activity!) can be done without the use of dietary supplements. Perhaps, this is the most important lesson to transmit to the public.
REFERENCES