Research claiming to show a benefit of a particular supplement often reaches the news prematurely. The fact that a supplement given in high doses to rats in a laboratory shows an improvement in an aspect of health suggests that further research should be carried out. Beware studies that use very small numbers of people over a very short period of time. Research should ideally take two large groups of people, give the supplement to one group and then observe who develops heart disease or cancer and who does not.
In reality, this would be impractical and prohibitively expensive.
Nutrition researchers rely instead on biomarkers to indicate the likelihood of developing disease. For example, heterocyclic amines are often used in colon cancer studies. These proteins significantly increase the risk of developing the disease. Changes in biomarkers allow researchers to suggest whether a supplement modifies risk. Very few researchers would state that supplement X will prevent cancer or heart disease; instead they talk about risk.
Even well-designed studies are difficult to interpret, as people have different lifestyles, genetic make-up and family history. We know how much of a nutrient is needed to prevent deficiency, but we do not yet know what levels would achieve optimum health. The only convincing evidence is that a healthy diet with plenty of fruit and vegetables can reduce the chances of dying from heart disease or cancer.
Dr Toni Steer is a nutritionist at the Medical Research Council's Human Nutrition Research unit, Cambridge